A feedback loop between heterochromatin and the nucleopore complex controls germ-cell to oocyte transition during Drosophila oogenesis

Summary

Germ cells differentiate into oocytes that upon fertilization launch the next generation. How the highly specialized oocyte acquires this distinct cell fate is poorly understood. During Drosophila oogenesis, H3K9me3 histone methyltransferase SETDB1 translocates from the cytoplasm to the nucleus of germ cells concurrent with oocyte specification. Here, we discovered that nuclear SETDB1 is required to silence a cohort of differentiation-promoting genes by mediating their heterochromatinization. Intriguingly, SETDB1 is also required to upregulate 18 of the ~30 nucleoporins (Nups) that comprise the nucleopore complex (NPC), promoting NPC formation. NPCs anchor SETDB1-dependent heterochromatin at the nuclear periphery to maintain H3K9me3 and gene silencing in the egg chambers. Aberrant gene expression due to loss of SETDB1 or Nups results in loss of oocyte identity, cell death and sterility. Thus, a feedback loop between heterochromatin and NPCs promotes transcriptional reprogramming at the onset of oocyte specification that is critical to establish oocyte identity.

Graphical Abstract

eTOC blurb

In Sarkar et al., the authors describe how a cohort of early oogenesis genes are silenced by heterochromatin formation during oocyte specification. The heterochromatin promotes nucleopore complex (NPC) formation that in turn helps maintain silenced genes to developmentally regulate gene silencing and fertility.

Introduction

Germ cells give rise to gametes that upon fertilization launch the next generation^1–3. Germ cells can become germline stem cells (GSCs) that self-renew and differentiate to give rise to sperm or an oocyte^1,4–6. The oocyte upon fertilization can differentiate into every cell lineage in the adult organism^7,8. The gene regulatory mechanisms that enable the transition from germ cells to oocytes are not fully understood.

Drosophila has a well-characterized transition from germline stem cell (GSC) to an oocyte^2,4. Drosophila ovaries comprise individual units called ovarioles that house the GSCs in the germarium (Figure 1A–A1)^7,9. GSC division results in a new GSC and a cystoblast, which differentiates via incomplete mitotic divisions, giving rise to 2-, 4-, 8- and 16-cell cysts (Figure 1A1)^10–12. One of these 16 cells is specified as the oocyte whereas the other 15 cells become nurse cells^13,14. Somatic cells envelop the nurse cells and the specified oocyte to form an egg chamber (Figure 1A1)⁹. The nurse cells produce maternal mRNAs, that are deposited into the specified oocyte mediated by an RNA binding protein, Egalitarian (Egl)^14,15. An inability to specify or maintain the oocyte fate leads to death of the egg chamber mid-oogenesis resulting in sterility^14,16.

Figure 1: SETDB1 and windei are required for silencing RpS19b reporter during oogenesis.

(A) A schematic of a Drosophila ovariole consisting of germarium and egg chambers surrounded by somatic cells (light red). Egg chambers grow and produce an egg (white).

(A1) A schematic of a Drosophila germarium. Germline stem cells (GSCs; light green) are proximal to somatic niche (red) and divide to give rise to daughter cells called cystoblasts (dark green). Both GSCs and cystoblasts are marked by spectrosomes (red). Cystoblasts differentiate giving rise to 2-, 4-, 8-, and 16-cell cysts (green), marked by fusomes (red). In the 16-cell cyst, one cell commits to meiosis and specifies an oocyte (dark blue) while the other 15-cells become nurse cells (light blue).

(B-B2) Confocal images of a germarium of a fly carrying RpS19b::GFP reporter transgene stained for GFP (green, right grayscale), Egl (red, right grayscale) and Vasa (blue). GFP is expressed in the undifferentiated stages and early cysts (white dashed line), while Egl is expressed in the differentiated cysts and localized to the specified oocyte (yellow arrows).

(C-G1) Ovariole of control RpS19b::GFP (C-C1), GKD of SETDB1 (D-E1) and wde (F-G1) stained for GFP (green, right grayscale), Vasa (blue) and 1B1 (red). Depletion of these genes resulted in egg chambers that ectopically express RpS19b::GFP (white dashed line), did not grow, and died mid-oogenesis (white solid arrows).

(H) Quantification of ovarioles with ectopic RpS19b::GFP expression upon GKD of SETDB1 or wde compared to control ovaries (N= 50 ovarioles; 96% in SETDB1 GKD #1, 92% in SETDB1 GKD #2 and 90% in wde GKD #1, 70% in wde GKD #2 compared to 0% in control). Statistics: Fisher’s exact test; *** = p<0.001.

(I) Arbitrary units (A.U.) quantification of RpS19b::GFP expression in the germarium and egg chambers upon GKD of SETDB1 (magenta) or wde (orange) compared to control ovaries (black). GFP is expressed in the undifferentiated cells and is attenuated in egg chambers. In SETDB1 and wde GKD, GFP expression persists in the egg chambers. Statistics: Dunnetťs multiple comparisons test; N= 10 ovarioles; ns = p>0.05, *=p<0.05, ** = p<0.01, *** = p<0.001.

Scale bars: 15 micron.

The transition from GSC to an oocyte requires dynamic changes in gene expression¹⁷. Once a GSC gives rise to the cystoblast, it expresses differentiation factor Bag of marbles (Bam), promoting its differentiation to an 8-cell cyst^18,19. In the 8-cell cyst, the expression of the RNA binding fox-1 homolog 1 (Rbfox1) is required to mediate transition into the 16-cell cyst stage, allowing for an oocyte to be specified²⁰. Translation of Rbfox1 requires increased levels of ribosomal small subunit protein 19 (RpS19) accomplished in part by expression of the germline specific paralog RpS19b in the undifferentiated and early differentiating stages^21,22. During differentiation, the germline also initiates meiotic recombination mediated by the synaptonemal complex consisting of proteins such as Sisters Unbound (Sunn), Corona (Cona) and Orientation Disruptor (Ord)^23–25. More than one cell in the cyst stage initiates recombination but as oocyte differentiation proceeds, only the specified oocyte retains the synaptonemal complex (Figure 1A1)^23,26,27. After oocyte-specification, the levels of mRNAs encoding RpS19b and some synaptonemal complex proteins are diminished, suggesting early oogenesis genes are no longer expressed²². How the expression of these early oogenesis genes is attenuated is not known.

In Drosophila, the SET Domain Bifurcated Histone Lysine Methyltransferase 1 (SETDB1) (also called Eggless) is required for deposition of gene silencing Histone H3 Lysine 9 trimethylation (H3K9me3) marks and heterochromatin formation^28–30. SETDB1 is expressed throughout Drosophila oogenesis, but as the oocyte is specified, it shifts from a cytoplasmic to predominantly nuclear localization³¹. A conserved cofactor called Windei (Wde) is required for either SETDB1 nuclear translocation or nuclear stability^32,33. Loss of SETDB1 during germline development results in an accumulation of undifferentiated cells^29,34. In addition, loss of SETDB1 and wde also result in egg chambers that do not grow and die mid-oogenesis^28,32. SETDB1 is known to be required for silencing transposons and male-specific transcripts in the female germline^29,34,35. However, neither the upregulation of transposons nor male-specific genes in female germline results in egg chambers that do not grow^36,37. Together these data suggest that SETDB1 silences a yet-unidentified group of genes to promote oogenesis.

Here, we find that genes that are expressed in early stages of oogenesis, including genes that promote oocyte differentiation and synaptonemal complex formation, are silenced upon oocyte specification, via a feedback loop between SETDB1-mediated heterochromatin and the nucleopore complex (NPC). Inability to silence these differentiation-promoting genes due to loss of either SETDB1 or members of the NPC results in loss of oocyte identity and death. Several aspects of germ cell differentiation have been studied and implicated in loss of fertility in sexually reproducing organisms. Our work indicates that a previously unappreciated broad transcriptional reprogramming silences critical aspects of the germ cell differentiation program at the onset of oocyte specification and is essential to promote oocyte identity.

Results

SETDB1 promotes silencing of RpS19b reporter at the onset of oocyte specification.

We hypothesized that the expression of early oogenesis mRNAs such as RpS19b is silenced upon oocyte specification. To monitor RpS19b expression, we used a reporter that expresses an RpS19b::GFP fusion protein from the endogenous RpS19b promoter. This RpS19b::GFP shows high expression in the germarium and attenuated expression post-oocyte specification and in the subsequent egg chambers, consistent with its endogenous RpS19b mRNA expression pattern (Figure 1B–C1, I)^22,38.

Using a previously characterized hemagglutinin (HA) tagged endogenous SETDB1, we found that a large fraction of SETDB1 translocates from the cytoplasm to the nucleus concurrent with oocyte specification (Figure S1A–A3)39. To test if SETDB1 is required for the silencing of RpS19b^28,31, we performed germline knockdown (GKD) of SETDB1, in the background of RpS19b::GFP reporter using two independent RNAi lines. We detected the germline, RpS19b::GFP, and spectrosomes/fusomes/somatic cell membrane in ovaries by immunostaining for Vasa, GFP, and 1B1, respectively^40,41. We found that, compared to the control, GKD of SETDB1 resulted in ectopic RpS19b::GFP protein expression in the differentiated egg chambers without affecting levels in the undifferentiated stages (Figure 1C–E1, H–I; Figure S1B). Thus, SETDB1 is required to repress RpS19b::GFP reporter in the differentiated egg chambers.

To determine if nuclear SETDB1 was required to repress RpS19b::GFP post-oocyte specification, we depleted wde, the cofactor required for SETDB1's nuclear localization, in the germline and independently assayed for SETDB1 nuclear localization and RpS19b::GFP (Figure S1C). GKD of wde resulted in loss of nuclear SETDB1 in the differentiated stages of oogenesis without affecting cytoplasmic levels in the undifferentiated stages (Figure S1D–F). We found that GKD of wde using two independent RNAi lines, like GKD of SETDB1, results in ectopic RpS19b::GFP protein expression in the egg chambers without affecting levels in the undifferentiated stages (Figure 1C–I). In addition to upregulation of RpS19b::GFP, GKD of both SETDB1 and wde resulted in egg chambers that did not grow in size and died mid-oogenesis as previously reported (Figure S1G)^28,32. Thus, repression of the RpS19b::GFP reporter in the differentiated egg chambers requires nuclear SETDB1.

To further analyze the requirement for SETDB1 and wde in silencing early oogenesis genes, we generated germline clones of the SETDB1 and wde using the FLP/FRT technique^32,42. As the absence of GFP expression marked germline clones, we could not use RpS19b::GFP expression as a read out. Instead, we used expression of another early oogenesis gene, blanks, as a readout^16,43. Blanks is a component of a nuclear siRNA pathway that has critical roles in the testis but does not have any overt function during oogenesis⁴³. While control clones showed no Blanks staining in the differentiated egg chamber (Figure S1H–H1), mutant clones for SETDB1 and wde showed Blanks expression in the differentiated egg chambers (Figure S1I–K). Our data suggests that nuclear SETDB1 is required for silencing early oogenesis genes such as RpS19b and blanks during oocyte differentiation.

SETDB1 and wde repress genes that are primarily expressed before oocyte specification.

To determine if SETDB1 and wde repress other differentiation-promoting genes in addition to RpS19b and blanks, we performed RNA Sequencing (RNA-seq). We compared ovaries from SETDB1- and wde-GKD flies to ovaries from control (UAS-Dcr2;nosGAL4) flies, including young virgin flies which lack late-stage egg chambers. Principal component analysis of the RNA-seq data revealed that SETDB1 and wde-GKD ovary transcriptomes closely resembles young virgin control rather than adult control (Figure S2A). Using a 1.5-fold cut off (Fold Change (FC)≥|1.5|) and False Discovery Rate (FDR)<0.05, we found that compared to young virgin control, 2316 genes were upregulated and 1972 were downregulated in SETDB1 GKD ovaries, and 1075 genes were upregulated and 442 were downregulated in wde-GKD ovaries (Figure 2A–B) (Supplemental Table 1). Moreover, comparing wde- to SETDB1-GKD ovaries showed significant overlap of the upregulated (80%) and downregulated (75%) transcripts, suggesting that SETDB1 and wde co-regulate a cohort of genes during oogenesis (Figure 2C; Figure S2B).

Figure 2: SETDB1/Wde represses a cohort of early oogenesis genes.

(A-B) Volcano plots of –Log₁₀P-value vs. Log₂Fold Change (FC) of (A) SETDB1 and (B) wde GKD ovaries showing significantly downregulated (pink) and upregulated (blue) transcripts in SETDB1, and wde GKD ovaries compared with control ovaries (FDR = False Discovery Rate < 0.05 and genes with 1.5-fold or higher change were considered significant).

(C) Venn diagram of upregulated genes from RNA-seq of SETDB1 and wde GKD ovaries compared to control. 862 targets are shared between GKD of SETDB1 and wde, suggesting that SETDB1 and Wde co-regulate a cohort of genes.

(D) The biological process GO terms of shared upregulated genes in ovaries depleted of SETDB1 and wde compared to control (Statistics: Fisher’s exact test), showing differentiation as one of the significant processes regulated by SETDB1/Wde.

(E-F) RNA-seq track showing that RpS19b and blanks is upregulated upon GKD of SETDB1 and wde.

(G) Violin plot of mRNA levels of the 862 shared upregulated targets in ovaries enriched for GSCs, cystoblasts, cysts, and whole ovaries, showing that the shared targets between SETDB1 and wde are expressed up to the cyst stages are attenuated in whole ovaries. Statistics: Hypergeometric test; *** indicates p<0.001.

(H-J1) Confocal images of germaria probed for RpS19b mRNA (red, grayscale) and DAPI (blue) in control (H-H1) showing RpS19b RNA expression restricted to germarium but in GKD of SETDB1 (I-I1) and wde (J-J1) ovarioles showing RpS19b mRNA expression is expanded to egg chambers.

(K-M1) Confocal images of germaria probed for blanks mRNA (red, grayscale) and DAPI (blue) in control (K-K1) showing blanks mRNA expression restricted to early stages of oogenesis and in GKD of SETDB1 (L-L1) and wde (M-M1) ovarioles where blanks mRNA expression is expanded to egg chambers.

(N-O) Quantification of fluorescence intensity of RpS19b (N) and blanks (O) mRNAs in the germarium and egg chambers depleted of SETDB1 (magenta) or wde (orange) compared to control ovaries (gray). Statistics: Dunnetťs multiple comparisons test; N= 10 ovarioles; ns = p>0.05, *=p<0.05, ** = p<0.01, *** = p<0.001.

Scale bars:15 micron.

SETDB1 and Wde are known to repress gene expression, thus we first focused on mRNAs with increased levels in the GKD ovaries^28,33. Gene Ontology (GO) analysis of the shared upregulated RNAs indicated that many were genes involved in differentiation (Figure 2D). Among the upregulated RNAs was RpS19b and blanks, validating our initial screen, as well as genes that promote synaptonemal complex formation such as sunn, ord and cona (Figure 2E–F; Figure S2C–E). Thus, SETDB1 and wde repress a cohort of RNAs that are either critical for differentiation or merely expressed during early oogenesis.

To determine when SETDB1 and Wde act to repress genes during oogenesis, we analyzed available RNA-seq libraries enriched for GSCs, cystoblasts, and cysts, early egg chambers and late-stage egg chambers²². We found that SETDB1/wde-regulated RNAs decreased after the cyst stages and their levels were attenuated in the later stages of oogenesis compared to non-targets (Figure 2G, Figure S2F–G)²². This reduction did not happen without SETDB1 and wde (Figure 2G). RNA in situ analysis of blanks, and RpS19b revealed that these mRNAs are present in the early stages of oogenesis and are attenuated after oocyte specification in controls but these RNAs persisted in SETDB1 and wde GKD egg chambers (Figure 2H–O). Thus, mRNAs broadly expressed before oocyte specification are repressed by SETDB1 and Wde in differentiated egg chambers.

SETDB1 represses transcription of a subset of targets by increasing H3K9me3 enrichment.

To investigate whether SETDB1/wde-regulated mRNAs are repressed at transcriptional level, we examined a subset of nascent transcripts (pre-mRNAs) by qRT-PCR. Indeed, the levels of nascent RpS19b, ord, sunn, cona and blanks mRNAs were increased in SETDB1/wde-GKD ovaries compared to control ovaries (Figure S3A–B). These data suggest that transcription of these genes increases upon loss of SETDB1 or wde. SETDB1 and its nuclear translocation by Wde is required for silencing genes expressed in the early stages of oogenesis. We found that GKD of SETDB1 reduced H3K9me3 throughout oogenesis, whereas GKD of wde reduced H3K9me3 in the differentiated egg chambers but not in the undifferentiated stages (Figure S3C–F). This suggested that SETDB1-mediated H3K9me3 heterochromatin is required to silence early oogenesis genes during oocyte differentiation.

To determine if the SETDB1-dependent repression of these genes involves changes in H3K9me3, we performed CUT&RUN^44,45 on adult control (UAS-Dcr2;nosGAL4) ovaries enriched for differentiated egg chambers where these genes are repressed (Figure 2G). Analysis of CUT&RUN data from adult control showed enrichment of H3K9me3 marks on previously identified SETDB1 targets and genes containing heterochromatin such as PHD Finger Protein 7 (phf7) and light (lt) respectively validating our CUT&RUN data (Figure 3A–B; Figure S3G)^34,46. As genes in the Drosophila genome are closely packed, we only analyzed the gene body from 5’UTR to the end of the 3’UTR to unambiguously identify SETDB1 regulated genes⁴⁷. We found that 1593 out of 2,316 genes upregulated upon loss of SETDB1 are enriched for H3K9me3 marks compared to IgG negative control (Figure 3C). In addition, we found that 888 genes lose H3K9me3 on their gene bodies upon GKD of SETDB1 including RpS19b and ATP-dependent chromatin assembly factor (Acf) (Figure 3D–G). The upregulated genes that do not show changes to H3K9me3 marks within the gene body may be regulated by elements outside of the gene body or indirectly. Our data suggest that SETDB1 is required for H3K9me3 enrichment and transcriptional repression of a cohort of early-oogenesis genes in the egg chamber.

Figure 3: SETDB1 promotes silencing of early oogenesis genes by regulating levels of H3K9me3.

(A-B) Tracks showing level of H3K9me3 on previously validated and known heterochromatic genes phf7 and lt respectively.

(C) Bar graph showing genes regulated by SETDB1 that are enriched for H3K9me3 on the gene body. 1593 (black) out of 2316 (gray) genes upregulated upon loss of SETDB1 are enriched for H3K9me3.

(D) Volcano plot showing changes in H3K9me3 in SETDB1 GKD compared to WT. 888 genes lose H3K9me3 after SETDB1 GKD (red).

(E) Bar graph showing 270 (pink) out of 886 (gray) genes that lose H3K9me3 enrichment in SETDB1 GKD were upregulated upon loss of SETDB1.

(F-G) Tracks showing level of H3K9me3 on target genes (top panel, red) and level of their transcripts in SETDB1 GKD (bottom panel, grey). Loss of H3K9me3 on SETDB1 targets RpS19b and Acf (G) respectively suggesting they are directly regulated by SETDB1.

(H-K1) Ovariole from control (H-H1) and GKD of SETDB1 (I-I1) in the background of UAS-GFP transgene. RNAi resistant WT SETDB1::GFP transgene (J-J1) and RNAi resistant Y→A point-mutant SETDB1::GFP transgene (K-K1) stained for GFP (green), H3K9me3 (red) and DAPI (blue). Depletion of SETDB1 results in loss of H3K9me3 expression which was rescued by the WT-SETDB1 transgene but not the Y→A point mutant. (100% in control, 97% in WT SETDB1::GFP and 0% in Y→A point mutant SETDB1 transgene; n=30).

(L-O1) Ovariole from control (L-L1) and GKD of SETDB1 (M-M1) in the background of UAS-GFP. RNAi resistant WT-SETDB1 transgene (N-N1) and RNAi resistant Y→A point mutant SETDB1 transgene in the background of SETDB1 GKD (O-O1) stained for GFP (green), Blanks (red) and DAPI (blue). Depletion of germline SETDB1 results in ectopic Blanks expression which was rescued by the WT but not in Y→A point-mutant transgene (100% in control, 93% in WT SETDB1::GFP and 3% in Y→A point mutant SETDB1 transgene, n=30)

Scale bars: 15 micron.

We next wanted to determine if catalytic activity of SETDB1 is required for its silencing of early oogenesis genes. The catalytic activity of SETDB1 family is mediated by a conserved tyrosine (Tyr, Y)^48,49. By aligning the sequences of SET domain from different species^48–50, as well as using the crystal structure of SET domain, we determined the identity of the catalytic tyrosine of Drosophila SETDB1 that promotes methyltransferase activity, to be Tyrosine 1050 (Tyr 30, when only the SET domain is considered)⁵⁰ (Figure S3H).

To determine if the catalytic function of SETDB1 is required to silence early oogenesis genes in the differentiated egg chambers, we created a germline specific expression (UASp) line to generate both a WT-SETDB1 (UAS-SETDB1^RNAi-res_WT-GFP) as well as a putative catalytically dead version of SETDB1 by replacing the predicted catalytic Tyr with Alanine (Y→A) mutation (UAS-SETDB1^RNAi-res_Y-A-GFP). Both WT-SETDB1 and putative catalytically dead SETDB1 were generated by a transgene that was re-coded via mutations at synonymous sites to make them resistant to RNAi knockdown. We then depleted endogenous SETDB1 in the germline using RNAi and then overexpressed both WT and putative catalytically dead version of SETDB1.

We stained for H3K9me3 under conditions when only WT or putative catalytically dead mutant was expressed in the germline, depleted of endogenous SETDB1. We found that the WT-SETDB1 transgene could rescue the endogenous depletion of SETDB1 and formed heterochromatin (Figure 3H–J1). In contrast, the Y→A mutant SETDB1 did not rescue the phenotype and was defective for heterochromatin formation (Figure 3H–K1). WT and the Y→A mutant SETDB1 were expressed in the germline and translocated to the nucleus during differentiation (Figure 3J–K1). Assaying for Blanks expression, we found that WT-SETDB1, but not the Y→A mutant SETDB1 transgene, silenced Blanks expression in the differentiated egg chambers (Figure 3L–O1). Our data suggests that Tyr 1050 contributes to catalytic activity of SETDB1 and is necessary to silence early oogenesis genes such as blanks.

SETDB1 is required for transposon repression during oogenesis^29,51, and the upregulation of transposons can affect gene expression^52,53. However, we found that the upregulation of genes in the differentiated stages that we observed upon depletion of SETDB1 was not due to the secondary effect of transposon upregulation as the expression of RpS19b reporter was not altered in germline depleted of aubergine (aub), a critical component of the piRNA pathway (Figure S3I–K)^35,36,54, nor did aub depletion cause mid-oogenesis death as we observed in SETDB1 and wde GKDs (Figure S3I–K)^54,55 . Overall, our data suggest that loss of SETDB1 derepresses a subset of genes during late oogenesis independent of transposon dysregulation.

SETDB1 is required for the expression of NPC components.

GO term analysis of downregulated targets of SETDB1/wde GKD included genes that regulate transposition, consistent with the previously described role of SETDB1/Wde in the piRNA pathway and those that regulate proper oocyte development, consistent with the previously described phenotype (Figure 4A)^29,31,32,51.

Figure 4: SETDB1/Wde promotes expression of a subset of nucleoporin genes and NPC formation.

(A) The significant biological process GO terms of downregulated genes in SETDB1 and wde GKD ovaries compared to control (FDR from p-values using a Fisher’s exact test), showing nucleocytoplasmic transport as one of the processes regulated by SETDB1/Wde.

(B) A schematic of the Nucleopore Complex (NPC) comprised of ~30 nucleoporins (nups) and organized into subcomplexes.

(C) Table showing levels of 18 nucleoporin mRNAs that are down regulated 1.5 or more-fold in both SETDB1 or wde GKD ovaries compared to control ovaries.

(D) qRT-PCR assaying the pre-mRNA levels of SETDB1 and Wde-regulated Nup genes, including Nup154, Nup205 and Nup107 are decreased compared to control while levels of non-target Nup62 pre-mRNA is not affected (control level vs SETDB1 GKD and wde RNA in=3, ** = p<0.01, *** = p<0.001, Error bars are SEM, Student’s t-Test).

(E-G3) Ovariole and egg chamber images of control (E-E3), GKD of SETDB1 (F-F3) and wde (G-G3) stained for NPC (red, grayscale), Vasa (green) and DAPI (blue). NPC staining was done using Mab414 antibody. Depletion of SETDB1 and wde shows reduced expression of NPC in the egg chambers suggesting SETDB1 regulates expression of several nucleoporins which in turn regulates formation of NPC.

(H-I) A.U. quantification of NPC level in the germline (H) and soma (I) in SETDB1 and wde GKD ovaries compared to control. Statistics: Dunnetťs multiple comparisons test; N= 25 ovariole for germline and 15 for somatic quantitation; ns = p>0.05, * = p ≤ 0.05, ** = p<0.01, *** = p<0.001.

Scale bars: 15 micron for main images, 4 microns for insets.

Unexpectedly, we observed that genes involved in nucleocytoplasmic transport were downregulated in SETDB1/wde-GKD ovaries compared to controls (Figure 4A). Nucleocytoplasmic transport is mediated by Nucleopore complexes (NPCs), which span the nuclear membrane and consist of a cytoplasmic ring, a central scaffold spanning the nuclear envelope, and a nuclear ring and basket (Figure 4B)^56–58. Beyond regulating nucleocytoplasmic transport, NPCs also regulate gene transcription by anchoring and maintaining heterochromatic domains^59–61. We found that GKD of SETDB1/wde in the germline resulted in downregulation of 18 out of ~30 nucleoporins (Nups) that make up the NPC (Figure 4C), including a germline enriched Nup154 that is critical for oogenesis^62–64. The Nups that were downregulated upon depletion of SETDB1 and wde were not isolated to one specific NPC subcomplex (Figure 4B–C).

We found that nascent mRNAs corresponding to the SETDB1/wde targets Nup154, Nup205 and Nup107 were downregulated in SETDB1/wde-GKD ovaries, whereas the non-target Nup62 was unaffected, suggesting that SETDB1/Wde promotes transcription of a cohort of Nups (Figure 4D). In addition, the levels of a Nup107::RFP fusion protein, under endogenous control⁶⁵, were significantly reduced in the cysts and egg chambers of SETDB1- and wde-GKD compared to controls (Figure S4A–D).

To determine if loss of Nup expressions in SETDB1/wde-GKD ovaries resulted in loss of NPC formation, we performed immunofluorescence with an antibody that is known to mark NPCs in Drosophila^56,66. We found that germline NPC levels were reduced in the egg chambers of SETDB1/wde-GKD ovaries compared to controls (Figure 4E–H), but the NPCs in the soma was unaffected (Figure 4I, Figure S4E–G2), and nuclear lamina were also unaffected (Figure S4H–K). Thus, SETDB1/wde are required for the expression of a subset of Nups and NPC formation after oocyte specification.

Heterochromatic genes and piRNA clusters require heterochromatin to promote their transcription^29,67. Although we found that SETDB1 is required for upregulation of Nups, CUT&RUN analysis of H3K9me3 marks revealed that only 3 of the Nup genes had any enrichment of H3K9me3 (Mbo, Nup188, Gp210). Moreover, among SETDB1-regulated Nups, only Gp210 showed any heterochromatic enrichment (Supplemental Table 2). Taken together, we find that SETDB1 indirectly promotes proper Nup expression by a yet unknown mechanism in the germline.

Nucleoporins are required to maintain heterochromatin domains at the nuclear periphery.

Our data indicate that, in Drosophila female germline, heterochromatin formation mediated by SETDB1 is required for proper NPC formation by promoting proper expression of a subset of Nups including Nup107 and Nup154 (Figure 4C). In yeast, a subset of Nups is part of the heterochromatin proteome and are required to cluster and maintain heterochromatin at the NPC^61,68,69. This subset includes Nup107 and the yeast homolog of Nup154, Nup155, which both have reduced expression in SETDB1/wde-GKD compared to controls. We hypothesized that in Drosophila, SETDB1 could promote silencing of early oogenesis genes by promoting heterochromatin formation. This heterochromatin then promotes expression of Nups and NPC formation, which can help maintain heterochromatin by anchoring it to nuclear periphery, thus promoting silencing of early-oogenesis genes.

To first determine if heterochromatin and nucleoporins associate in Drosophila female germline, we utilized antibody against H3K9me3 to mark heterochromatin and Nup107::RFP to mark NPCs in control ovarioles (nosGAL4; Nup107-RFP)^29,65. We found that H3K9me3 domains were often at the nuclear periphery, proximal to Nup107::RFP (Figure 5A–A2, C). Next, to determine if loss of Nups leads to loss of heterochromatin, we first depleted Nup154 and probed for heterochromatin formation. We chose Nup154, as its loss of function phenotype has been well described^63,64. We found that GKD of Nup154 in the germline, resulted in egg chambers that do not grow and die mid-oogenesis as previously described for Nup154 mutants (Figure S5A–C)⁶³. In addition, upon depletion of Nup154 translocation of SETDB1 from the cytoplasm to the nucleus, monitored by immunostaining, was not affected, suggesting that attenuation of heterochromatin upon GKD of Nups is not due to loss of transport of SETDB1 into the nucleus (Figure S5D–F). By staining for H3K9me3 marks, we found that upon GKD of Nup154, heterochromatin domains initially form (Figure S5G–H3). However, in the egg chambers of Nup 154 GKD, the colocalization between H3K9me3 domains and Nup107::RFP levels at the nuclear periphery were significantly reduced before reduction of heterochromatin levels (Figure 5A–C, Figure S5G–H3). GKD of Nup107 also resulted in egg chambers that do not grow and lose heterochromatin (Figure S5F–K). Moreover, loss of germline Nup154 results in increased distance between heterochromatin domain and the nuclear periphery marked by LaminC (Figure 5D–F). Thus, Nups 154 and 107, positively regulated by SETDB1, are required for H3K9me3 localization at the nuclear periphery for H3K9me3 maintenance in the female germline.

Figure 5: H3K9me3 heterochromatin colocalizes with NPC component Nup107 at the nuclear periphery.

(A-A2) Egg chambers of control ovariole showing RFP-Nup107 (red, right red channel), H3K9me3 (green, right green channel). Heterochromatin is seen in close association with NPC (white arrows). Colocalized fraction is shown in yellow.

(B-B2) Egg chambers of Nup154 GKD ovariole showing significant decrease in the colocalization (white arrows) between RFP-Nup107 (red, right red channel) and H3K9me3 (green, right green channel).

(C) Quantification of fraction of H3K9me3 that colocalizes with NPC in the germline of control ovarioles (gray) in contrast to Nup154 GKD ovarioles (blue). Quantitative object based colocalization was measured in Imaris software, *** = p<0.001, one-tailed Students t-Test.

(D-E) Egg chambers of control (D) and Nup154 GKD (E) ovariole showing significant increase in the distance between LamC (red) and H3K9me3 (green). Single nuclei from control and Nup154 GKD ovariole are shown in the insets.

(F) Quantification of distance between H3K9me3 and LamC in the germline of control ovarioles (gray) in contrast to Nup154 GKD ovarioles (blue). Statistics: *** = p<0.001, one-tailed Welch’s t-Test.

Scale bars: 15 micron.

Nups are required for silencing early-oogenesis genes.

Based on our findings above that Nups are required to maintain H3K9me3 levels and localization, we hypothesized that they are also required to silence the early-oogenesis RNAs in differentiated egg chambers. To test this hypothesis, we depleted Nup154 and Nup107 in the germline of a fly carrying the RpS19b::GFP reporter. We found that GKD of these nucleoporins resulted in upregulation of RpS19b::GFP phenocopying GKD of SETDB1/wde (Figure 6A–C; Figure S6A–B1, D). Moreover, germline depletion of Nup62, which is within the NPC but not regulated by SETDB1, also resulted in upregulation of RpS19b::GFP and egg chambers that did not grow (Figure S6A–D). This suggests that activity of NPC components and not just the Nups regulated by SETDB1 are required for silencing RpS19b::GFP reporter.

Figure 6: Nup154 is required for silencing a cohort of genes expressed during early oogenesis.

(A-B1) Ovariole of control RpS19b::GFP (A-A1), GKD of Nup154 (B-B1) stained for GFP (green, right grayscale), Vasa (blue) and 1B1 (red). Depletion of Nup154 results in ectopic expression of RpS19b::GFP (white dashed line) and egg chambers that did not grow.

(C) Arbitrary units (A.U.) quantification of RpS19b::GFP expression in the germarium and egg chambers upon GKD of Nup154 (blue) compared to control (gray). GFP is expressed in the germarium and then attenuated upon egg chamber formation in control. In Nup154 GKD, GFP expression persists in the egg chambers. Statistics: Dunnetťs multiple comparisons test; N= 10 and 8 ovarioles for control and Nup154 GKD respectively; ns = p>0.05, *=p<0.05, ** = p<0.01, *** = p<0.001.

(D) Volcano plots of –Log₁₀P-value vs. Log₂Fold Change (FC) of mRNAs that show significantly downregulated (pink) and upregulated (blue) transcripts in Nup154 GKD ovaries compared with control ovaries (FDR = False Discovery Rate < 0.05 and 1.5-fold or higher change were considered significant).

(E) Venn diagram of upregulated overlapping genes from RNA-seq of SETDB, wde and Nup154 GKD ovaries compared to control. 751 upregulated targets are shared between SETDB1, wde and Nup154 GKD, suggesting that Nup154 and SETDB1 function to co-regulate a cohort of genes.

(F) RNA-seq track showing that RpS19b is upregulated upon germline depletion of Nup154.

(G) Violin plot of mRNA levels of the 2809 upregulated targets in ovaries enriched for GSCs, cystoblasts, cysts, and whole ovaries, showing that the upregulated targets of Nup154 are most highly enriched up to the cyst stages, and then attenutated in whole ovaries. Statistics: Hypergeometric test; *** indicates p<0.001.

(H) Venn diagram showing overlapping genes that lose H3K9me3 after depletion of both SETDB1 and Nup154 in the germline. 622 genes lose H3K9me3 after Nup154 GKD out of which 564 genes are also directly silenced by SETDB1, suggesting co-regulation of these genes by both SETDB1 and Nup154.

(I-J2) Ovariole of control (I-I2), GKD of Nup154 (J-J2) probing for RpS19b genomic locus (red, right grayscale) by DNA in situ, and stained for LamC (green, right grayscale) and DAPI (blue). Depletion of Nup154 shows increased distance of RpS19b locus from nuclear periphery in nurse cells (white arrows). One nurse cell nucleus is shown with a yellow dotted circle for both control and Nup154 GKD.

(K) Quantitation of distance between RpS19b locus from the nuclear periphery in nurse cells of control (gray) and Nup154 GKD (blue) in micron. Distance was measured in ImageJ using the straight-line function. Statistics: ns = p>0.05, *=p<0.05, ** = p<0.01, *** = p<0.001, Welch’s t-Test.

Scale bars: 15 micron.

To determine if Nups are required for silencing other early oogenesis RNAs, we performed RNA-seq, and compared Nup154 GKD ovaries with young ovaries (UAS-Dcr2;nosGAL4) as a developmental control (Figure S2A). Using a 1.5-fold cut off (Fold Change (FC)≥|1.5|) and False discovery rate (FDR)<0.05), we found that compared to control, in Nup154 GKD 2809 genes are upregulated, and 2922 genes are downregulated (Figure 6D) (Supplemental Table 1). Strikingly, 97% of upregulated genes and 89% of downregulated SETDB1/wde targets overlapped with Nup154 GKD (Figure 6E; S6E). Nup154 was involved in silencing genes that promote oocyte differentiation, including synaptonemal complex components ord, sunn and cona, and RpS19b (Figure 6F; S6F–H). In addition, GKD of Nup154 also resulted in upregulation of blanks (Figure S6I). The levels of Nup154-regulated RNAs decreased after the cyst stage, when the oocyte is specified, in contrast to non-targets, which have similar RNA levels at all stages (Figure 6G; S6J–K). Thus, Nup154 is critical for silencing early-oogenic mRNAs in the differentiated egg chambers.

To determine if Nup154 is required for H3K9me3 marks at SETDB1-regulated gene loci such as RpS19b, we carried out CUT&RUN for H3K9me3 in control and Nup154 GKD. We found that 564 out of 622 genes displaying a loss in H3K9me3 in Nup154 GKD also show the same loss in SETDB1 GKD including RpS19b and Acf (Figure 6H; S6L–M) (Supplemental Table 2). We find that Nups are required for silencing and maintaining H3K9me3 at a subset of SETDB1/wde-regulated loci.

To ascertain if loss of proper NPC formation results in mis-localization of heterochromatin targets like RpS19b from the nuclear periphery, we performed DNA in situ hybridization of the gene locus using RpS19b DNA probes^70–72. In situ hybridization and staining for nuclear lamina with LaminC antibody showed the presence of this gene locus proximal to nuclear lamina in control (Figure 6I–I2). However, loss of Nup154 in the germline resulted in increased distance between RpS19b locus and nuclear lamina (Figure 6I–K). This suggests that post oocyte specification, proper formation of NPC is required to maintain RpS19b locus at the nuclear periphery.

NPC loss affects H3K9me3 heterochromatin but not H3K27me3 heterochromatin.

H3K27me3 can also promote silencing during Drosophila oogenesis^73,74. NPC is required to maintain SETDB1 deposited H3K9me3 heterochromatin by anchoring it to the nuclear periphery. To determine if NPC also promotes H3K27me3 gene silencing marks we stained for H3K27me3 in both SETDB1 and Nup154 GKD. We found that H3K27me3 level increased upon the germline loss of SETDB1 and Nup154 (Figure S7A–D). This suggests that the effects of SETDB1 and Nup154 are specific to H3K9me3 but not H3K27me3 repressive marks.

Silencing genes expressed during the early oogenesis stages is required for maintaining oocyte fate.

We next asked why loss of SETDB1, wde and Nups results in egg chambers that do not grow and die mid-oogenesis. Egg chambers with oocyte specification or maintenance defects result in death of egg chambers mid-oogenesis¹⁶. To determine if there are oocyte specification or maintenance defects, we stained GKD of SETDB1, wde and Nup154 for the oocyte marker Egalitarian (Egl) as well as Vasa and 1B1^14,15. In the early stages of oogenesis, as in control, GKD of SETDB1, wde and Nup154 resulted in Egl seemingly localizing to one cell (Figure 7A–E). However, in the later egg chambers, compared to control ovariole, GKD of SETDB1, wde and Nup154 resulted in either mis-localization or diffused Egl expression suggesting loss of oocyte fate (Figure 7A–E). Thus, SETDB1, wde and Nup154 are required to maintain an oocyte fate.

Figure 7: Silencing of early oogenesis genes mediated by SETDB1, Wde and Nup154 is required for maintenance of oocyte fate.

(A-D1) Ovarioles of control (A-A1), GKD of SETDB1 (B-B1), wde (C-C1) and Nup154 (D-D2) stained for Egl (green, right grayscale), Vasa (blue) and 1B1 (red). Control shows proper oocyte specification while depletion of SETDB1, wde and Nup154 in the germline results in initial oocyte specification (yellow arrow) which is then lost in the subsequent egg chambers (yellow dashed line).

(E) Quantification of percentage ovarioles with abnormal/loss of Egl expression (black) in ovaries depleted of SETDB1 or wde or Nup154 compared to control ovaries (gray) (N= 50 ovarioles; 98% in SETDB1 GKD and 100% in wde and Nup154 GKD compared to 0% in control.) Statistics: Fisher’s exact t-test. *** = p<0.001.

(F) A model showing that nuclear SETDB1 after differentiation promotes heterochromatin formation. This heterochromatin promotes NPC formation which in turn helps maintain heterochromatin.

Scale bars: 15 micron.

Discussion

Many maternally contributed mRNAs in oocytes are critical for early development after fertilization^{13,14,75–77}. We previously showed that many mRNAs expressed in germ cells and the undifferentiated stages of oogenesis must be selectively degraded and thus excluded from the maternal contribution¹⁶. However, the potential role of transcriptional silencing of germ cell and GSC-enriched genes during oogenesis was unclear. Here, we found that regulated translocation of SETDB1 into the nucleus during oocyte specification is required to silence germ cell- and early oogenesis-genes in the differentiated egg chambers (Figure 7F), and that this process is essential to maintain oocyte fate. Thus, some genes expressed in germ cells and some that promote differentiation are transcriptionally silenced at the onset of oocyte specification mediated by a feedback loop between heterochromatin and NPC.

Regulated heterochromatin formation during oocyte specification promotes germ cell to oocyte transition.

A large fraction of SETDB1 is cytoplasmic in the undifferentiated stages of the germline. As the oocyte is specified during differentiation, SETDB1 becomes mostly nuclear²⁸. This translocation of SETDB1 to the nucleus during oocyte specification is mediated by Windei (Wde), the Drosophila ortholog of mAM/MCAF1^32,33. Here we find that translocation of SETDB1 to the nucleus during oocyte specification is required to silence germ cell and early-oogenesis genes at the onset of oocyte specification. MCAF1 also regulates the accumulation of SETDB1 in the nucleus in mammalian cells⁷⁸. In addition, loss of SETDB1 during mammalian oogenesis results in meiotic defects and infertility⁷⁹. These data suggest that regulated heterochromatin formation to promote silencing of early oogenesis genes could be conserved to regulate oogenesis in mammals.

We discovered that SETDB1 is required to silence two major classes of genes. The first group is involved in GSC differentiation into an oocyte, including critical genes that promote meiosis I. The second group of genes are those that are expressed in the germ cells before differentiation into an oocyte but have no specific function in the female germline such as blanks^16,43. We propose that these genes silenced upon oocyte specification are detrimental to late oogenesis or early embryogenesis. Indeed, it has been shown that overexpression of one such gene actin 57B (act57B), which is repressed by SETDB1/Wde (Supplemental Table 1), is detrimental to oogenesis^16,80. Remarkably, some of the mRNAs encoded by genes that SETDB1 transcriptionally silences during this transition are also targeted at the post-transcriptional level for degradation by members of the no go decay pathway such as blanks and Act57B¹⁶. Thus, our data suggests that the regulation of gene expression during oocyte differentiation reflects a two-step process: transcriptional silencing dependent on SETDB1, and post-transcriptional degradation of mRNAs to exclude a cohort of germ cell mRNAs from the maternal contribution¹⁶.

Nucleopore complex and heterochromatin are in a feedback loop to promote gene silencing.

The NPC not only mediates selective nucleo-cytoplasmic transport of macromolecules but also regulates gene expression by anchoring chromatin domains, including heterochromatin to the nuclear periphery^61,81,82. In addition, several Nups are also part of the heterochromatin proteome in yeast, suggesting that NPCs can regulate gene expression by regulating heterochromatin^61,83. Consistent with these observations, we find that NPC and heterochromatin are closely associated in the female germline of Drosophila. Loss of NPCs due to depletion of individual Nups results in loss of heterochromatin and subsequent upregulation of germ cell and early oogenesis genes resulting in oogenesis defects. The large overlap of target genes between SETDB1, wde and Nup154 is indicative that Nups are functioning in the same pathway as SETDB1. This suggests that not only do NPCs associate with heterochromatin, but that NPCs also play a role in maintaining heterochromatin and gene repression during oocyte differentiation.

The number of genes that need to be silenced varies based on cell types and developmental trajectory. How levels of heterochromatin are coupled to their NPC docking sites in the cell was not known. Like heterochromatin levels, the number of NPCs varies by cell type and during differentiation⁸⁴. How NPC number is regulated during development was not fully understood. Our findings in the female germline suggest an elegant tuning mechanism for heterochromatin and its NPC docking sites. Heterochromatin promotes levels of NPC which then promote heterochromatin maintenance by tethering it to the nuclear periphery. We find that this loop can be developmentally regulated by controlling levels of SETDB1 in the nucleus mediated by conserved protein Wde to promote heterochromatin formation.

Limitations of the study

We do not know how SETDB1 is guided to its targets to promote their silencing. Similarly, what triggers SETDB1 translocation from the cytoplasm to nucleus during oocyte specification promoting heterochromatinization of the early oogenesis genes is not known. Loss SETDB1 results in upregulation of a large set of early oogenesis genes but we find that a cohort of these upregulated genes are the direct targets with heterochromatin on their gene body such as RpS19b. We do not know how SETDB1 controls silencing of targets such as blanks. Lastly, we find that SETDB1 promotes expression of Nups and we also find that Nup genes are not heterochromatic genes. We speculate that SETDB1 regulates NPC formation indirectly, but the mechanism is not known. Finally, it is possible that multiple Nups but not all Nups are required in silencing early oogenesis genes during oocyte differentiation.

STAR Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Prashanth Rangan (prashanth.rangan@mssm.edu).

Materials availability

All flies generated and used in this study are available on request.

Data and code availability

RNA-seq and CUT&RUN data have been deposited at Gene Expression Omnibus (GEO). Data generated during this study are available at GEO databank under accession number: GSE186982.

Any additional information required to reanalyze the data reported in this work paper is available from the Lead Contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Husbandry conditions of experimental animals

Flies were grown at 25–29°C and dissected between 0–3 days post-eclosion.

Fly food was made using the procedures as previously described (summer/winter mix) and narrow vials (Fisherbrand Drosophila Vials; Fischer Scientific) were filled to approximately 10–12mL⁹².

Fly lines

The following RNAi stocks were used in this study; if more than one line is listed, then both were quantitated and the first was shown in the main figure: SETDB1 RNAi (Perrimon lab) (Rangan et al., 2011²⁹), Bloomington #24106), Wde RNAi (Bloomington #33339, VDRC #105719), Nup154 RNAi (Bloomington #34710), Nup62 RNAi (Bloomington #35695), Nup107 RNAi (Bloomington #43189), Nup205 RNAi (VDRC #V38608), FRT42B, SETDB1¹⁴⁷³ (Wodarz lab), FRT42B, wde^TD63 (Wodarz lab). Germ line clones for SETDB1 and wde were generated using a heat shock promoter driven flippase on the X-chromosome³².

The following tagged lines were used in this study: dSETDB1-HA (Bontron Lab)³⁹, RpS19b::GFP (Buszczak Lab, ²²), mRFP-Nup107 (Bloomington #35516), UAS-EGFP (Bloomington #5431), UAS-SETDB1^RNAi-res_WT-GFP (This study), UAS-SETDB1^RNAi-res_Y-A-GFP (This study).

The following tissue-specific drivers and double balancer lines were used in this study: UAS-Dcr2;nosGAL4 (Bloomington #25751), nosGAL4;MKRS/TM6 (Bloomington #4442), and If/CyO;nosGAL4 (Lehmann Lab).

METHOD DETAILS

UAS-SETDB1-GFP overexpression line construction

We re-coded the sequence targeted by the SETDB1 RNAi line by making synonymous changes at each codon based on the codon usage table in Drosophila. The synonymous changes made the gene resistant to RNAi knockdown but did not alter the amino-acid sequence of SETDB1. Using the RNAi resistant line as background we also designed a mutant line where the predicted catalytic Tyrosine residue was re-coded such that the mutant gene codes for Alanine instead. The resulting insert sequence was synthesized by GenScript and cloned into a plasmid we generated that contains UASp to drive them in the germline and has GFP to mark the presence of the transgene. We submitted the transgenic constructs to The BestGene Inc (Chino Hills, CA) for injection into w¹¹¹⁸ flies.

Dissection and Immunostaining

Ovaries were dissected and teased apart with mounting needles in cold PBS and kept on ice. All incubation was done with nutation. Samples were fixed for 10 minutes in 5% methanol-free formaldehyde. Ovaries were washed in 0.5 mL PBT (1X PBS, 0.5% Triton X-100, 0.3% BSA) 4 times for 5 minutes each. Primary antibodies in PBT were added and incubated at 4°C nutating overnight. Samples were next washed 3 times for 5 minutes each in 0.5 mL PBT, and once in 0.5 mL PBT with 2% donkey serum (Sigma) for 15 minutes. Secondary antibodies were added in PBT with 4% donkey serum and incubated at room temperature for 3–4 hours. Samples were washed 3 times for 10 minutes each in 0.5 mL of 1X PBST (0.2% Tween 20 in 1x PBS) and incubated in Vectashield with DAPI (Vector Laboratories) for 1 hour before mounting.

The following primary antibodies were used: mouse anti-1B1 (1:20; DSHB), Rabbit anti-Vasa (1:1,000; Rangan Lab), Chicken anti-Vasa (1:1,000; Rangan Lab) (Upadhyay et al., 2016), Rabbit anti-GFP (1:2,000; abcam, ab6556), Rabbit anti-H3K9me3 (1:500; Active Motif, AB_2532132), Mouse anti-H3K27me3 (1:500; abcam, ab6002), Rabbit anti-Egl (1:1,000; Lehmann Lab), Mouse anti-NPC (1:2000; BioLegend, AB_2565026) and Rat anti-HA (1:500; Roche, 11 867 423 001). The following secondary antibodies were used: Alexa 488 (Molecular Probes), Cy3 and Cy5 (Jackson Labs) were used at a dilution of 1:500. For each stainings, we have stained at least 5 pairs of ovaries and each experiment was repeated 3 times independently.

Fluorescence Imaging

The tissues were visualized, and images were acquired using a Zeiss LSM-710 confocal microscope under 20X, 40X and 63X oil objective with pinhole set to 1 airy unit. All gain, laser power, and other relevant settings were kept constant for any immunostainings being compared. Image processing was done using Fiji and gain adjustment and cropping was performed in Photoshop CC 2019.

Egg laying assays

Assays were conducted in vials with 3 control or experimental females under testing and 1 wild type control males. Crosses were set up in triplicate for both control and experimental. All flies were 1-day old post-eclosion upon setting up the experiment. Cages were maintained at 29°C and plates were changed daily for counting. Analyses were performed for 5 consecutive days. Number of eggs laid were counted and averaged. Adult flies eclosed were counted from all the vials and averaged.

RNA isolation

Ovaries from flies were dissected in cold 1x PBS. RNA was isolated using TRIzol (Invitrogen, 15596026)^16,22.

RNA was treated with DNase (TURBO DNA-free Kit, Life Technologies, AM1907), and then run on a 1% agarose gel to check integrity of the RNA.

RNA-seq library preparation and analysis

Libraries were prepared using the Biooscientific kit. To generate mRNA enriched libraries, total RNA was treated with poly(A)tail selection beads (Bioo Scientific Corp., NOVA-512991). Manufacturer’s instructions of the NEXTflex Rapid Directional RNA-seq Kit (Bioo Scientific Corp., NOVA-5138-08) were followed, but RNA was fragmented for 13 minutes. Library quality was assessed with a Fragment Analyzer (5200 Fragment Analyzer System, AATI, Ankeny, IA, USA) following manufacturer's instructions. Single-end mRNA sequencing (75 base pair reads) was performed on biological duplicates from each genotype on an Illumina NextSeq500 by the Center for Functional Genomics (CFG).

After quality assessment, the sequenced reads were aligned to the Drosophila melanogaster genome (UCSCdm6) using HISAT2 (version 2.1.0) with the RefSeq-annotated transcripts as a guide⁸⁶. Differential gene expression was assayed by DeSeq2, using a false discovery rate (FDR) of 0.05, and genes with 2-fold or higher were considered significant. The raw and unprocessed data for RNA-seq generated during this study are available at Gene Expression Omnibus (GEO) databank under accession number: GSE186982 (Token number: wlenykcoldmzfqf). GO term enrichment on differentially expressed genes was performed using Panther⁹⁴.

Fluorescent in situ hybridization

A modified RNA in situ hybridization procedure for Drosophila ovaries was followed. RNA probes were designed and generated by LGC Biosearch Technologies using Stellaris^® RNA FISH Probe Designer, with specificity to target base pairs of target mRNAs. Ovaries (3 pairs per sample) were dissected in RNase free 1X PBS and fixed in 1 mL of 5% formaldehyde for 10 minutes. The samples were then permeabilized in 1mL of Permeabilization Solution (PBST+1% Triton-X) rotating in RT for 1 hour. Samples were then washed in the wash buffer for 5 minutes (10% deionized formamide and 10% 20x SSC in RNase-free water). Ovaries were covered and incubated overnight with 1ul of probe in hybridization solution (10% dextran sulfate, 1 mg/ml yeast tRNA, 2 mM RNaseOUT, 0.02 mg/ml BSA, 5x SSC, 10% deionized formamide, and RNase-free water) at 30°C. Samples were then washed 2 times in 1 mL wash buffer for 30 minutes and mounted in Vectashield.

For DNA in situ hybridization, the DNA probes were generated by the Joyce lab at University at Pennsylvania. A modified DNA in situ hybridization procedure for Drosophila imaginal discs was followed. Ovaries were dissected and teased in 1X PBS and fixed in 800ul of fixative solution (10% formaldehyde, 1ul 100% NP-40, 20ul 10X PBS) for 10 minutes. The ovaries were quickly washed and permeabilized in PBX (1.5 mL Triton-X in 500 mL 1X PBS) for 30 minutes. Samples were then incubated overnight with primary antibodies in PBX at 4°C. Ovaries were then washed with 1X PBX and incubated with secondary antibodies in 1X PBX for 2 hours at RT. The samples were quickly washed with 2X SSCT (500ul Tween-20 in 50 mL of 2X SSC). The samples were then subjected to 3 consecutive washes with 2X SSCT + 20% formamide, 2X SSCT + 40% formamide, 2X SSCT + 50% formamide (vol/vol) for 10 minutes each at RT. the last wash was repeated twice. Then the ovaries were transferred to a PCR tube in 50% formamide and DNA was pre-denatured at 37°C for 4 hours, 92°C for 3 minutes, and 60°C for 20 minutes. The formamide solution was then replaced with 36ul of probe buffer (50% formamide in 2X SSCT + 10% dextran sulfate + 4ul of probe) + 1ul with RNase A. Avoid adding more than 4ul of probe in one reaction mixture. This was followed by overnight incubation (19 hours) at 37°C in the dark with shaking. Next, the samples were washed twice with 50% formamide-2X SSCT solution at 37°C with shaking for 30 minutes each followed by one 10 minutes wash in 20% formamide-2X SSCT solution at RT. The ovaries were then mounted in vectashield.

CUT&RUN assay

Ovaries from flies were dissected in ice cold 1x PBS and ovarioles were separated by teasing after dissection with mounting needles. PBS was removed and the samples were permeabilized in 1mL of Permeabilization Solution (PBST+1% Triton-X) rotating in RT for 1 hour. Samples were then incubated overnight at 4°C in primary antibody dilutions in freshly prepared BBT+ buffer (PBST + 1% BSA + 0.5 mM Spermidine + 2 mM EDTA + 1 large Roche complete EDTA-free tablets). Primary antibody was replaced with BBT+ buffer and quickly washed twice. Samples were then incubated in ~700 ng/ml of pAG-MNase in BBT+ buffer rotating for 4 hours at 25°C. Samples were then quickly washed twice in wash+ buffer (20 mM HEPES pH7.5 + 150 mM NaCl + 0.1% BSA + 0.5 mM Spermidine + 1 large Roche complete EDTA-free tablets in water). Samples were resuspended in 150 μl Wash+C (wash+ + 100 mM CaCl₂) and incubated for 45 minutes on nutator at 4°C. The cleavage reaction was terminated by addition of 150 μl StopR (NaCl final 200 mM + EDTA final 20 mM + 100μg/mL RNaseA) and incubating the sample at 37˚C for 30 minutes. Samples were then centrifuged at 16,000 x g for 5 minutes and 300 μl of the supernatant was collected for DNA discovery. To the supernatant, 2 μL 10% SDS and 2.5 μL of 20 mg /mL Proteinase K was added and incubated at 50°C for 2 hours. Half of this was kept as a backup and half was used in bead cleanup. 20 μL AmpureXP bead slurry and 280 μL MXP buffer (20% PEG8000 + 2.5 M NaCl + 10 mM MgCl2 in water) was added to the sample and mixed thoroughly followed by 15 minutes incubation at RT. The beads were separated by magnet and supernatant was discarded. The beads were carefully washed with 80% ethanol for 30 seconds, while on the magnetic stand and air dried for 2 minutes. The beads were then resuspended in 10 μL DNase free water.

DNA seq library preparation and analysis

The samples from CUT&RUN assay were used for library preparation using NEBNext^® Ultra^™ DNA Library Prep Kit for Illumina^® (E7645, E7103) and adaptor ligated DNA were prepared without size selection.

CUT&RUN data analysis

CUT&RUN libraries were sequenced as paired-end 75bp reads on the Illumina NextSeq 500 at the University at Albany Center for Functional Genomics. FASTQ files were aligned to the dm6 reference genome using HISAT2 (10.1038/s41587-019-0201-4) (-X 10 -I 1000 –no-spliced-alignment, --no-discordant). Mapping statistics and data will be available from Gene Expression Omnibus. Alignment files were sorted and indexed using samtools and were subsequently used to create bigwig files for visualization with deeptools (--binSize 10)⁹¹. Principal component analysis between samples was performed using the multiBigwigSummary and plotPCA modules from deeptools. Only gene bodies were considered, and problematic genomic regions (blacklist) were removed from the analysis⁹⁵. Raw read counts of H3K9me3 enrichment across gene bodies was calculated using the HOMER annotateRepeats function and differential enrichment was calculated using DESeq2⁹⁶. H3K9me3 occupied genes are those with differential enrichment of H3K9me3 compared to IgG matched control conditions using DESeq2.

Quantitative Real Time-PCR (qRT-PCR)

1 μL of cDNA from each genotype was amplified using 5μL of SYBR green Master Mix, 0.3 μL of 10μM of each reverse and forward primers in a 10 μL reaction. The thermal cycling conditions consisted of 50°C for 2 minutes, 95°C for 10 minutes, 40 cycles at 95°C for 15 seconds, and 60°C for 60 seconds. The experiments were carried out in technical triplicate and minimum 2 biological replicates for each sample. To calculate fold change in mRNA levels, comparison was done to rp49 mRNA levels which was used as the control gene. Average of the 2^ΔCt for the biological replicates was calculated. Error bars were plotted using standard error of the ratios and P-value was determined by Studenťs t-test.

QUANTIFICATION AND STATISTICAL ANALYSIS

Quantifications of egg chamber area and fluorescent intensity

To quantify antibody staining intensities for GFP, RFP, HA, NPC, LamC, H3K9me3 and H3K27me3 or in situ probe fluorescence, images for both control and experimental ovariole were taken using the same confocal settings. Z stacks were obtained for all images. Similar planes in control and experimental were chosen, the area of cells or nuclei positive for the proteins or in situs of interest was outlined and quantified using the ‘measurement’ tool in Fiji (ImageJ). The mean intensity and area of the specified region was obtained. An average of all the ratios (mean/area), for the proteins or in situs of interest, per image was calculated for both control and experimental. Germline intensities were normalized to somatic intensities, if the protein or in situ of interest is germline enriched and not expressed in the soma, they were normalized to Vasa.

GFP levels in RpS19b::GFP flies were measured by outlining undifferentiated cells, cysts and egg chambers in germline specific manner and mean GFP intensity was measured. Nuclear HA levels in SETDB1-HA flies were quantified by outlining nucleus of cells pre and post oocyte specification and measuring mean HA intensity of a z stack. Cytoplasmic HA level was measured by quantifying the mean level of HA of a constant cytoplasmic area in multiple cells. HA level was normalized to either DAPI or somatic HA level in the follicle cells. H3K9me3 level was measured by outlining the nucleus and measuring the mean intensity of H3K9me3. It was then normalized to DAPI level. We also measured H3K9me3 intensity using stacks (see below). For the NPC quantification, we outlined the nuclear envelope based on the DAPI channel and measured the intensity of the NPC (mab414) channel. The NPC level quantified includes protein level at the nuclear envelope and not in the nucleoplasm or cytoplasm. Same was done for quantifying RFP and LamC in germline and somatic cells. Germline RFP/LamC intensity was then normalized to surrounding somatic level.

To measure the area of the germlines of egg chambers, planes in the middle of the egg chamber were chosen for control and experimental ovariole and put into a stack. Next, using the vasa channel the germline of each egg chamber was outlined and area was measured using ‘measurement’ tool in ImageJ.

For all measurements, a minimum of 5 pairs of ovaries were dissected and a minimum of five independent ovarioles were used for all quantitation. For all quantification, N represents number of ovarioles, which is mentioned in respective figure legends.

GraphPad (Prism) was used for all statistical analysis, details of which can be found in the figure legends. No method was used to determine whether the data met assumptions of the statistical approach. The statistical analysis for the data is found in the figures and figure legends.

Colocalization analysis

Confocal images of control and Nup154-RNAi mutants labeled for RFP-Nup107, H3K9me3, and DAPI were imported into Bitplane Imaris 9.6.2 for 3D reconstruction and colocalization analysis. Colocalization between RFP-Nup107 and H3K9me3 was calculated on a per egg chamber basis using the Surface-surface coloc function of Imaris and an automatic threshold detection and the surface-to-surface coloc function. The number of colocalized voxels was then normalized to the number of H3K9me3 voxels⁹³.

Quantification of H3K9me3 mark distance from nuclear periphery

Confocal images of control and Nup154 GKD mutants, labeled for LaminC and H3K9me3 were imported into Bitplane Imaris 9.6.2 for 3D spatial analysis. Nuclei of interest were isolated and the LaminC label defining the nuclear periphery was used to generate surface objects.

The center of mass for each H3Kme3 puncta was identified using the spots object function in Imaris. The distance of each H3K9me3 spot to the closest LaminC surface boundary was generated and the shortest distance was calculated (spot to surface distance). The mean of the distances of all H3K9me3 spots in a nucleus to the closest laminin surface boundary was calculated for both control and Nup154 GKD mutants. One-tailed Welch’s t-Test was used to calculate p-values for significance.

Supplementary Material

1

Supplementary Table1: Differential expression analysis from RNAseq of ovaries depleted of SETDB1, wde and Nup154 GKD compared to a developmental control, related to Figure 2 and Figure 6. Sheet1: Content: The table of contents for all the sheets in the spreadsheet.

2

Supplementary Table2: DEseq2 output using all H3K9me3 CUT&RUN datasets (WT, SETDB1 GKD, Nup154 GKD) versus IgG controls to determine H3K9me3-enriched regions, related to Figure 3 and Figure 6. Sheet1: Content: The table of contents for all the sheets in the spreadsheet.

4

Supplementary Table3: List of primers used in this study related to STAR methods.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-GFP	abcam	ab6556
Mouse monoclonal anti-GFP	Sigma	G6539
Rat monoclonal anti-HA high affinity Mouse anti-1B1	Developmental Studies Hybridoma Bank	Antibody Registry ID: 528070
Rabbit polyclonal anti-Vasa	Rangan Lab	N/A
Chicken polyclonal anti-Vasa	Rangan Lab	N/A
Rabbit polyclonal anti-Blanks	Gift from Sontheimer Lab	N/A
Rabbit polyclonal anti-Egalitarian	Gift from Lehmann Lab	N/A
Rabbit polyclonal anti-H3K9me3	Active Motif	39062
Mouse monoclonal anti-H3K9me3	Active Motif	61013
Mouse monoclonal anti-H3K27me3	abcam	Ab6002
Mouse monoclonal anti-NPC (mAb414)	BioLegend	902902
Mouse monoclonal anti-LamC	Developmental Studies Hybridoma Bank	Antibody Registry ID: AB_528339
Anti-rabbit Alexa 488	Jackson ImmunoResearch Labs	Code:711-546-152
Anti-mouse Alexa 488	Jackson ImmunoResearch Labs	Code: 715-546-150
Anti-mouse Cy3	Jackson ImmunoResearch Labs	Code: 715-166-150
Anti-rabbit Cy3	Jackson ImmunoResearch Labs	Code: 711-166-152
Anti-rat Cy3	Jackson ImmunoResearch Labs	Code:712-166-150
Anti-chicken Alexa 647	Jackson ImmunoResearch Labs	Code:703-606-155
Chemicals, peptides, and recombinant proteins
Formaldehyde (Methanol Free), 10% Ultrapure	Polysciences Inc.	#04018-1
Donkey Serum	Sigma-Aldrich	SKU: D9663
Vectashield Antifade Mounting Medium with DAPI	Vector Laboratories	#H-1200
Triton X-100 detergent	VWR	#97062-208
Nonidet P-40 (NP-40) substitute	IBI Scientific	#9016-45-9
Tween-20 detergent	VWR	#97062-332
TRIzol	Invitrogen	#15596026
Complete, EDTA-free Protease Inhibitor Cocktail Pill	Sigma-Aldrich	SKU: 5892953001
OmniPur® Formamide, Deionized	Calbiochem	4650
Pierce™ 16% Formaldehyde (w/v), Methanol-free	Thermo Fisher Scientific Inc.	28906
10X PBS buffer, pH7.4	invitrogen	AM9625
UltraPure 20X SSC buffer	Invitrogen	15557-044
BSA	VWR	E588-100G
SuperScript II	Invitrogen	18064022
RNase A	Thermo scientific	EN0531
Proteinase K	Thermo scientific	EO0491
EDTA	Millipore Sigma	CAS No.: 6381-92-6
99% Spermidine	Beantown Chemical	215885-1G
HEPES	Thermo Fisher Scientific Inc.	CAS No.: 7365-45-9
Polyethylene Glycol 8000	Millipore Sigma	CAS No.: 25322-68-3
Critical commercial assays
TURBO DNA-free Kit	Life Technologies	AM1907
NEXTflex Rapid Illumina DNA-Seq Library Prep Kit	BioO Scientific	NOVA-5138-11
NEBNext® Ultra II DNA Library Prep Kit for Illumina	New England Biolabs Inc.	NEB #E7645, E7103
Stellaris® RNA FISH Hybridization Buffer	LGC Biosearch Technologies	SMF-HB1-10
SYBR Green Master Mix	Applied Biosystems	4367659
Deposited data
RNA seq Data	This study	GSE186982
CUT&RUN Data	This study	GSE186982
Experimental models: Organisms/strains
UAS-Dcr2;nosGAL4	Bloomington Drosophila Stock Center	25751
nosGAL4;MKRS/TM6	Bloomington Drosophila Stock Center	4442
UAS-EGFP	Bloomington Drosophila Stock Center	5431
If/CyO;nosGAL4	Lehmann lab	N/A
SETDB1 RNAi	Perrimon lab; Bloomington Drosophila Stock Center	N/A; 24106
Wde RNAi	Bloomington Drosophila Stock Center; Vienna Drosophila Resource Center	33339; V105719
Nup154 RNAi	Bloomington Drosophila Stock Center	34710
Nup62 RNAi	Bloomington Drosophila Stock Center	35695
Nup107 RNAi	Bloomington Drosophila Stock Center	43189
Nup205 RNAi	Vienna Drosophila Resource Center	V38608
FRT42B, SETDB1 1473	Wodarz lab	N/A
FRT42B, wde TD63	Wodarz lab	N/A
dSETDB1-HA	Bontron Lab	N/A
RpS19b::GFP	Buszczak Lab	N/A
mRFP-Nup107	Bloomington Drosophila Stock Center	35516
UAS-SETDB1 RNAi-res_WT -GFP	This study	N/A
UAS-SETDB1 RNAi-res_Y-A -GFP	This study	N/A
Oligonucleotides
Primers	This study TableS3	N/A
Stellaris Probe Against blanks labeled with CALFluor590	LGC Biosearch Technologies	CDS
Stellaris Probe Against RpS19b labeled with CALFluor590	LGC Biosearch Technologies	CDS
DNA probes against RpS19b gene locus	Gift from Joyce lab	N/A
Recombinant DNA
Gateway Destination Vector Plasmid: pPGW	Drosophila Genomics Resource Center	Gateway 1 Collection
Software and algorithms
ImageJ	(Schindelin et al., 20128)	https://imagej.nih.gov/ij/
HISAT2	(Kim et al., 201586)	https://ccb.jhu.edu/software/hisat2/index.shtml
DESeq2	(Love et al., 201487)	http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html
featureCounts	(Liao et al., 201488)	http://bioinf.wehi.edu.au/featureCounts/
ggplot2	(Wickham, 201689)	https://cran.r-project.org/web/packages/ggplot2/index.html
FastQC	(Andrews, 201090)	https://www.bioinformatics.babraham.ac.uk/projects/fastqc/
deepTools	(Ramírez et al., 201691)	https://deeptools.readthedocs.io/en/develop/

Highlights.

• H3K9me3 heterochromatin silences early oogenesis genes during oocyte specification

• H3K9me3 heterochromatin is required for nucleopore complex formation

• Function of NPCs is required for silencing the early oogenesis genes

• Silencing of early oogenesis genes is essential for maintenance of oocyte fate