Introduction

Epithelial tissue fusion is a crucial process for building epithelial tubular network and uniting separate ends [1, 2]. Studies from different model systems, like neural tube fusion and Drosophila tracheal development, have provided several mechanistic insights into epithelial fusion process. For instance, specialized cells at the leading edge migrate and interact using cytoskeleton protrusions [3-9] and small GTPases like Rac and Cdc42 [10]. Upon contact, the leading cells re-establish polarity and connections with the help of extracellular matrix molecules, apical proteins and E-Cad [11]. However, the exact regulators and mechanisms controlling epithelial fusion remain to be determined, especially in a context-specific manner.

The reproductive system in animals comprises tissues from different origins, and must unite to form a connected system for producing and delivering gametes. For example, the human testis is directly linked to a complex tubular system consisting of the efferent and epididymal ducts [12]. Anomalies in the fusion of these ducts are a significant cause of cryptorchidism, affecting at least three percent of newborn boys. However, there is little understanding of the molecular mechanisms guiding the patterning and fusion of these ducts.

Similarly, the Drosophila male RS includes the gonad-derived testis and the genital disc (GD)-formed seminal vesicle, accessory gland (AG), ejaculatory duct (EjD) and additional supportive structures. During development, the TE and SV grow toward each other and merge into a continues tubular system [13]. This involves two epithelial ends, the terminal epithelial cells in the testis and the SV epithelium. While terminal epithelial cells in the testis are specified already in the embryo as male-specific somatic gonadal precursors (msSGPs) [14]. The SV epithelium, together with AG, develops from a group of mesenchymal cells in the GD which undergo mesenchymal-to-epithelial transition (MET) during the larval stage (Fig. 1A) [15, 16]. In the developing RS, multiple events, including epithelial tube formation and fusion as well as myoblast and pigment cell migration, are tightly coordinated (Fig. 1B) [17]. Thus, the developing Drosophila RS provides a representative in vivo context to dissect mechanisms underlying epithelial tissue assembly.

Figure 1.

Ham belongs to the PR domain containing protein family whose members are evolutionarily conserved cell fate regulators [18, 19]. Dysfunction of the human orthologs of Ham, PRDM16 (Mel1) and PRDM3 (Evi1, MECOM), has been found in a variety of human abnormalities including congenital diseases, metabolic disorders and cancers [20-27]. Moreover, in at least two types of epithelial cancers (kidney and pancreas cancers), PRDM16 was shown to play a tumor suppressive role [24, 27], indicating its importance in maintaining epithelial integrity. However, it remains undermined how these PR factors regulate epithelial tissue development and homeostasis.

Ham was exclusively studied in the nervous system where it is required for neuronal dendrite morphogenesis [28] and cell fate maintenance [29-31]. Ham was shown to repress gene expression through modulating chromatin states [32] or interacting with the co-repressor CtBP [29, 32]. However, the molecular functions of PRDM16 and PRDM3 are versatile, with their PR domains possessing intrinsic histone methyltransferase activity [21, 33] and the ability to interact with various co-activators and co-repressors [18, 34]. It remains unknown whether Ham plays a role outside of the nervous system and whether Ham also has dual function in gene expression.

In this study, we discovered that Ham is a master regulator in guiding epithelial tissue patterning and fusion in the Drosophila RS. Even a slight reduction of Ham disrupts epithelial differentiation and tissue assembly in the RS, ultimately affecting sperm and egg delivery. To explore new mechanisms governing epithelial tissue fusion in the RS, we conducted systematic analyses using RNA-sequencing (RNA-seq), Cleavage Under Targets and Tagmentation (CUT&TAG), multiplexed in situ hybridization data and genetic screening. These analyses led to identification of gene regulatory networks required for epithelial tissue fusion. Ham activates a gene network to promote epithelial cell differentiation and remodeling while suppressing another gene network involved in cell specification and early patterning. These target genes exhibit spatial-temporal expression dynamics during the fusion of the TE and SV. As a key Ham downstream effector, Wnt2 signaling regulates epithelial cell communication and interaction. By elucidating the regulatory network controlled by Ham, our study provides novel insights into the molecular mechanisms underlying epithelial tissue fusion.

Results

Drosophila RS development is vulnerable to reduced Ham activity

While PRDM16 and MECOM are critical cell fate regulators in various developmental contexts in mammals, the genetic roles of Ham were mainly found in the central and periphery nervous systems [28-31]. To systematically study Ham function, we generated new frameshift alleles using CRISPR/Cas9. The sgRNA targets the third exon shared by the three long isoforms B, D and G (Supplementary Fig. 1A), producing new alleles specifically disrupting the three PR-domain containing isoforms. We therefore named them ham^PRΔ. The original ham¹ and the null allele ham^SK1 shifted open reading frame (ORF) for all predicted isoforms [28, 30].

We validated nucleotide deletions in the new alleles by sanger sequencing (Supplementary Fig. 1B) and protein deletion by western-blot using an antibody against Ham. The Ham antibody detected several bands in the blot (Supplementary Fig. 1C). The two bands at and above the 150KD marker position in the two heterozygotic controls were not detectable in the ham^SK1 homozygous mutant, suggesting that these two bands correspond to the longest and the second longest isoforms respectively. The ham^PRΔ homozygous samples lack the largest band, verifying the depletion of the longest isoforms. We used one of them (-5nt) for all the later experiments.

We compared viability and sensory organ development of these alleles. In contrast to the original report showing that ham mutant animals died at the larval stage [28], we found that ham^SK1, ham^PRΔ and ham¹ alleles all survived to hatch when combined in trans with the ham deficiency line (Df) which deletes the entire ham locus (Supplementary Fig. 1D). While ham¹/Df and ham^SK1/Df animals all died right after hatching, about two thirds of the ham^PRΔ/Df mutants were viable without obvious gross defects. Furthermore, Ham is known to specify internal cell types (the neuron and sheath) in the sensory organ lineage, as ham¹mutants exhibited ectopic external cell types (the socket and shaft) with the expense of internal cells [29]. ham^PRΔ mutants displayed normal sensory organs, in contrast to defective sensory organs in the ham¹ and ham^SK1 alleles (Supplementary Fig. 1E). Thus, the PR-containing isoforms of Ham are dispensable for sensory organ development and viability.

However, ham^PRΔ mutant animals showed impaired fertility and abnormal morphology in both male and female RS (Fig. 1D-E, Supplementary 2A-B). The TE, AG and SV structures were disrupted in males, while mature eggs were stuck in the mutant female ovaries, indicating failed egg delivery. To confirm that the defects are induced by the ham mutation, we performed a rescue experiment by utilizing a pBAC transgene that carries the genomic region of ham with the super-folding GFP (sfGFP) coding sequence in frame with the ham ORF. The pBAC transgene rescued the viability of all ham mutant conditions, and at least partly restored male fertility of these mutants. Together, these results indicate that the ham^PRΔ allele represents a hypomorphic condition which preserves Ham shorter isoforms and partial activity. Thus, although viability and sensory organ development do not require the PR-containing isoforms, formation of the RS depends on them.

Ham-expressing epithelial cells connect the gonadal and supportive tissues

To examine whether Ham directly controls RS development, we examined Ham expression in the developing RS using a V5 antibody which detects all Ham isoforms in the ham_V5 allele [35]. Ham_V5 was detected in the connecting tube of the adult SV and TE (Fig. 1F). Correspondingly, at the late L3 stage Ham_V5 was detected in the terminal epithelial cells of the TE as well as the btl-Gal4-expressing mesenchymal cells in the genital disc which will give rise to the SV and AG (Fig. 1G-H). At the pupal stage, a high level of Ham persists in the TE terminal epithelial cells, SV and AG. In contrast, Ham is barely detectable in the SV myoblasts labeled by the reporter, mef2-Gal4-driving UAS-GFP (Fig. 1I-J). Similarly, in the female, high levels of Ham_V5 were found at the joint sites of oviducts and ovaries during both pupal and adult stages (Supplementary Fig 2C). Thus, Ham is an intrinsic factor in epithelial cells at the connection site between the gonadal- and genital-disc derived tissues in both males and females.

Ham is required for epithelial tissue fusion in the developing male RS

The specific expression of Ham and mutant phenotypes in the RS prompted us to investigate its function in RS epithelial tissues. Here we focused on the male, as the assembly of the TE and SV represents a remarkable process involving a series of tissue reorganization events (Fig. 1A-C). When the two ends of TE and SV get close, the epithelial cells at the leading edges meet and form a common epithelial tube. In the meanwhile, myoblasts from the SV and pigment cells from the TE migrate toward opposite directions, eventually forming distinct cell layers covering both structures.

We found that at 40-50h APF the TE and SV tubes in ham^PRΔ/Df and hamSK1/Df mutants did not properly elongate and appeared shorter and wider compared to that in the wild-type (Fig. 2A). The SV and AG were even unseparated in the ham null allele hamSK1/Df, suggesting that the severity of RS abnormality correlates with the levels of reduction in ham gene products.

Figure 2.

To examine epithelial integrity, we performed immunofluorescent staining using the antibodies against the epithelial cell markers, E-Cad (adherent junctions) and Coracle (Cor, septate junctions equivalent to mammalian tight junction) protein. We also incorporated an SV-specific transgenic reporter, btlGal4-driving UAS-EGFP or UAS-CD8GFP, to the control and ham mutant genotypes. E-Cad is enriched at the apical side (the lumen side, Fig. 2B) of the TE and SV epithelial tubes. Notably, the TE terminal epithelium accumulates a higher level of E-Cad compared to the SV epithelium, while Cor is mainly expressed at the SV side but not much in the TE, suggesting a heterotypic nature of the TE terminal and SV epithelia.

Control animals always showed connected TE and SV, whereas ham mutant TE and SV tissues were often separated from each other. And even in some mutants where TE and SV tubes were in close contact, they were disconnected (Fig. 2B). Importantly, this phenotype can be fully rescued by the pBAC_ham_sfGFP transgene, confirming that it is caused by the ham mutation. These observations suggest that the epithelial ends of the TE and SV possess distinct properties despite both being epithelial cell types and that Ham is required for TE and SV fusion.

Depletion of Ham abolishes TE terminal epithelial cell formation

A close inspection on the pupal TE revealed that most of TE epithelial cells were missing in ham mutant animals (Fig. 2B). TE terminal epithelial cells derive from male specific gonadal precursors (msSGPs) specified in the early embryo (Supplementary Fig. 3B) [15] and expand during larval and early pupal stages. ham mutant lava either lacked TE epithelial cells or showed abnormal location of these cells, which can be rescued by the pBAC_ham_sfGFP transgene (Supplementary Fig. 3A). This result indicates that Ham regulates specification or survival of these cells. Indeed, we detected Ham expression in msSGPs and other SGPs in the embryonic gonad. The number of msSGPs labeled by Abd-B or Sox100B-GFP was dramatically reduced in S13 ham mutant embryo (Supplementary Fig. 3C-G). Thus, Ham is a master regulator for the specification and/ or survival of msSGPs, and the absence of msSGPs accounts for loss of TE terminal epithelial cells in larval and pupal stages.

Ham regulates patterning and cell differentiation of the RS epithelial tubes

The epithelial cells in the SV and AG originate from a group of mesenchymal cells recruited by the FGF pathway involving Btl and Branchless (Bnl) (Fig. 1A and 1H) [15, 16]. These cells undergo MET to form an epithelial mass which then cavitates to generate inner hollow space and eventually elongates into an epithelial tube. In comparison with the control, the signal intensity of both E-Cad and Crumbs (Crb) were significantly reduced in ham mutant SV epithelial cells, which coincides with the shortening of septate junctions labeled by Cor (Fig. 2C-E). This result suggests that Ham regulates epithelial cell differentiation in the SV. To check whether Ham regulates epithelial cell differentiation in the TE, we used the FRT/FLP system (see Methods) to generate mosaic somatic mutant clones. In these animals, ham mutant cells were GFP negative, while wild-type and ham heterozygous cells expressed GFP (Fig. 2F). The E-Cad levels at the border between two mutant cells were significantly lower than those between wild-type cells, indicating that ham depletion impaired differentiation of TE terminal epithelial cells.

In summary, these phenotypic analyses suggest that Ham controls formation of TE terminal epithelial cells in the early embryo and regulates epithelial tube elongation and cell differentiation at the pupal stage. Thus, failed epithelium fusion in ham mutant RS could result from: 1) absence of TE terminal epithelial cells, 2) defects in epithelial tube elongation and cell differentiation of the SV and 3) a combination of both.

Ham directly controls epithelial fusion between the TE and SV

Next, we sought to assess whether and how Ham directly regulates the fusion process. We first attempted to generate a mild mutant condition which can bypass defects of TE epithelial cell loss and SV-AG tube elongation. We chose RNAi knockdown because knockdown efficiency can be enhanced by overexpressing Dicer2 (Dcr2) and modulated by temperature when RNAi is controlled by the GAL4/UAS system.

We first identified a Gal4 driver expressed in the SV and AG epithelial cells, after screening the available Gal4 transgenes driven by a series of genomic fragments spanning the ham locus [36] (Supplementary Fig. 4A). This transgene (thereafter named as hamRSGAL4) induces high expression of UAS-GFP in the TE and SV epithelial cells and neglectable expression in the myoblasts (Supplementary Fig. 4B). hamRSGAL4, when driving a ham RNAi line, nearly abolished Ham expression in animals carrying the UAS-dcr2 transgene (Fig. 2G-H, the middle panels). These animals also had a reduced number of TE epithelial cells and displayed a ball-like structure of the GD, similar to the ham null allele. Without UAS-dcr2, ham RNAi knockdown reduced Ham expression but preserved TE epithelial cells and normal-appearing SV. Remarkably, the TE and SV epithelia still failed to fuse (Fig. 2H, the right panel), suggesting that Ham also controls the late step of fusion. Furthermore, E-Cad staining intensity was significantly reduced in ham RNAi TE but not in ham RNAi SV (Fig. 2I-J), suggesting that expression of E-Cad in the TE is sensitive to even mild reduction of Ham protein levels.

Ham controls distinct molecular pathways to promote tissue assembly

To identify mechanisms that control epithelial fusion, we started to investigate Ham downstream target gene network. We performed RNA-seq experiments from dissected TE and SV tissues during the stage of 18-25h APF. We used two pairs of genotypes, including the wild-type (wt), and ham^PRΔ homozygous and the wt/Df and hamSK1/Df trans-heterozygous (Fig. 3A). The comparison between the first pair will determine dysregulated genes in the ham^PRΔ hypomorphic condition, while the comparison between the second pair will identify genes that change expression between the null and ham heterozygous conditions.

Figure 3.

To not exclude small gene expression changes (e.g. in the leading cells which are the minority within the whole tissue), we used a loose cutoff (FDR < 10% and FC>1). We found more than two thousand up-regulated and down-regulated genes in the GD. In the mutant TE, there were similar numbers of dysregulated genes (Fig. 3B). The genes that only changed expression in one condition could result from unspecific genetic background, differential dependence on the PR-containing isoforms or different sensitivity to ham heterozygous condition.

Expression of many dysregulated genes can be indirectly affected by Ham. We next intended to identify genes that are transcriptionally regulated by Ham. To this end, we determined Ham genomic binding sites using Cleavage Under Targets and Tagmentation assays (CUT&TAG) [37] on finely-dissected TE and grossly-dissected abdominal tissues in which Ham is mainly expressed in the TE and SV epithelium (Fig. 3A). We used a Ham antibody for wildtype tissues, a GFP antibody for tissues from the pBAC_ham_SFGFP transgene and IgG as a negative control in parallel for all the samples. Using the SEACR software, we identified peaks and intersections of peaks across the Ham and GFP antibody conditions (see Methods). For subsequent analyses, only the common 3433 and 691 peaks identified by the two antibodies (Ham and GFP) were included to ensure signal specificity (Fig. 3C-D).

To determine Ham target genes, we intersected the RNA-seq data from the hamSK1/Df versus wt/Df comparison with the Cut&Tag data. Among the 2,546 up-regulated genes in the GD, 441 (17.3%) contained Ham CUT&TAG peaks. Among the 2379 downregulated genes, 569 (23.9%) showed Ham binding (Fig. 3E). In the TE, a lower fraction of dysregulated genes contains Ham CUT&TAG peaks, including 4.7% (108/2274) upregulated genes and 7% (144/2057) downregulated genes (Fig. 3F), suggesting that dysregulated genes in the mutant TE are mainly indirectly controlled by Ham.

Both up- and down-regulated genes contain Ham binding peaks, suggesting that Ham can repress and activate gene expression. To systematically explore the function of Ham-regulated genes, we performed gene network analysis using STRING [38]. We focused on the genes bound by Ham and also dysregulated in the genital disc (GD) in the null condition hamSK1, considering that most of TE genes are only indirectly influenced in the ham^SK1 mutant. The SRING MSL clustering analysis generated 114 functional clusters from down-regulated genes and 62 clusters from up-regulated genes (Fig. 4A, Supplementary Table 1). Interestingly, many of downregulated genes clustered into several large groups including cytoskeleton regulators, metabolic factors, cell adhesion and extracellular matrix proteins, suggesting that Ham activates genes to promote epithelial differentiation and remodeling. In contrast, upregulated genes are enriched for patterning factors, cell cycle genes and transcriptional and translational regulators, suggesting that Ham suppresses early developmental processes. Notably, metabolic factors are enriched in both categories, but these factors belong to distinct functional groups. Thus, epithelial fusion requires complex and balanced metabolic processes. Together, this network analysis confirms that Ham positions on top of distinct molecular pathways to facilitate TE and SV assembly.

Figure 4.

Ham was known for its repressive activity via interacting with the co-repressor CtBP and influencing histone modifications [32]. Our finding of its activation function was unexpected. To validate the activation activity of Ham, we chose three downregulated target genes (Fig. 4B), dwnt2, shot and shg, for transcriptional reporter assays. shg encodes E-cad, while Shot is a member of the spectraplakin family of large cytoskeletal linker molecules simultaneously binding to actin and microtubules. The function of wnt2 was reported in the developing RS development [39, 40]. We cloned four genomic fragments from the wnt2 locus, two each from shg and shot into a luciferase construct, and co-transfected each reporter with individual plasmids expressing several variants of Ham. In addition to the Ham long isoform and a PR truncation version of Ham, we included the NZF and the CZF domains fused with the VP16 activation domain (NZF_VP16 and CZF_VP16) (Fig. 4C). The two ZF domains of Ham are putative DNA binding domains, but it remained undetermined whether they help to recruit Ham to DNA. This experimental setting allows to assess which of the ZF domains is able to target Ham-regulated enhancers. The full-length, the PR truncation and the NZF_VP16 fusion all activated expression of the reporters driven by 7 out of 8 Ham-bound elements (Fig. 4D-G, Supplementary Fig. 5A-B). By contrast, the CZF_VP16 had little activity on the reporters. This result confirms that Ham can directly activate target gene expression and that the NZF domain directly or indirectly recruits Ham to enhancers.

To further identify co-factors with which Ham interacts to regulate gene expression, we performed de novo motif discovery from Ham CUT&TAG peaks in mixed TE and GD tissues close to up-regulated and down-regulated genes in the hamSK1 mutant using the MEME-ChIP software. This resulted in nine and five significantly enriched motifs from up- and down-regulated gene loci respectively (Supplementary Fig. 5D). While five motifs were shared between these two sets, the motifs for Dpp/BMP signaling factor MAD, the zinc-finger protein ZNF263 and two unknown factors were only enriched in the upregulated genes. The result implies that Ham may switch activation to repression activity when interacting with these co-factors in the context of RS development.

Ham downstream effector genes are required for male RS fertility

Built upon our transcriptomic data, we next intended to identify genes crucial for the fusion process. We reasoned that as the majority of TE terminal epithelial cells are lost in both ham mutant alleles, the commonly downregulated genes in ham^PRΔ and ham^SK1TE (Fig. 3B) may be expressed in the wt TE epithelial cells and can be and functionally important. We thus focused on the top 241 common downregulated genes whose expression showed a fold change greater than 1.5, and the three validated Ham target genes: shot, wnt2 and shg. Using all available RNAi lines from the Bloomington stock center that cover 113 out of 244 candidate genes, we performed a UAS-RNAi screen with the UAS-dcr2; hamRSGal4 driver for male fertility (see Methods).

The RNAi screen led to identification of 26 genes whose knockdown abolished or reduced male fertility (Supplementary Table 2). Knocking down four of these genes, shg, AP2-alpha, pebble (pbl) and CG11406, resulted in complete sterility, indicating their importance for male fertility. We defined this category as 100% penetrance of sterility. Knockdown of other twenty-two genes reduced fertility to different extent (Fig. 5A-B).

Figure 5.

Consistent with the gene network analysis (Fig. 4A), these genes include cytoskeleton regulators (pbl, shg, shark) [41-43] and metabolism factors (CG11406, Ph4alphaMP, GstE11, CG17477, CG11069, AdSL, CG17283, CG15117, CG7484). Other genes are known or predicted to be involved in intracellular membrane trafficking (AP-2alpha, CG31935 and CG7456) [44] and signaling (wnt2, Toll, AkhR, Idgf4 and Hr46) [45-48].

We confirmed that knocking down pbl, AP-2alpha, CG11406 and shg led to failed TE and SV fusion and abnormal morphology at the pupal stage (Fig. 5C). pbl encodes a Rho guanine nucleotide exchange factor, implicated in multiple processes involving cytoskeleton reorganization[41, 49, 50]. The gene product of CG11406 is a predicted triglyceride lipase, cleaving triacylglycerol (TAG) to produce diacylglycerol (DAG). AP-2alpha, a subunit in the Adaptor complex 2, is involved in intracellular trafficking and secretion [51]. Remarkably, shg knockdown also led to aberrant mobilization of TE epithelial cells to the middle of the TE, while no such effect was found in the SV and AG. Downregulation of E-Cad is a key event during epithelial to mesenchymal transition (EMT), and E-Cad promotes cohesive cell migration of mesenchymal cells, but suppressing cell motility of epithelial cells [52, 53]. Our result suggests that the SV and the TE terminal epithelia have distinct dependence on E-Cad and that TE epithelial cells are prone to becoming motile when E-Cad is reduced. Together, this screen identified a collection of known and novel factors involved in epithelial tissue fusion, confirming that the fusion process requires coordinated activity of several classes of molecules.

Ham downstream genes exhibit spatial-temporal expression dynamics in the male RS

During epithelial fusion, cells at the leading edge become specialized and express specific genes required for tissue interaction and remodeling [1, 2]. Moreover, as we observed, the TE terminal and SV epithelia have distinct properties, suggesting that these two sets of epithelial cells may express distinct genes. We thus intended to assess spatial distribution of Ham downstream genes. We conducted a high-resolution multiplexed RNA in situ hybridization assay known as SCRINSHOT (single cell RNA in situ hybridization on tissues) [54] to visualize expression of candidate genes including wnt2, shg, shot and other positive genes from the RNAi screen, along with the receptor genes in Wnt2 signaling (fz, fz2, fz3 and fz4) and ligand genes in Tl-signaling (spz, spz4, spz5 and spz6) (Supplementary Table 3). Some of these Wnt2 and Tl signaling genes displayed dysregulation in at least one of the ham mutant alleles according RNA-seq data (Supplementary Fig 6A-B). Other receptor genes in Wnt signaling and ligand genes in Tl signaling were omitted due to low read numbers in the RNA-seq data, indicating negligible or minimal expression in the TE and GD.

Each SCRINSHOT reaction took place within a mini-chamber, accommodating pairs of control and ham mutant tissues, or GFP KD control and ham RNAi KD samples (refer to Methods for details of probe design and experimental procedures). This setup was designed to minimize technical variability between the genotypes. mRNAs are visualized as fluorescent dots in SCRINSHOT images, with each dot representing a single mRNA molecule. However, due to the densely packed nature of cells in the TE and SV tissues, quantifying mRNA molecules at a single-cell resolution is challenging. Therefore, we opted to quantify fluorescent signal intensity within defined areas (see Methods). In total, 37 genes were tested, including ham itself (Supplementary Table 3). Probes for 7 genes gave either no or low signal in the TE or GD, likely attributed to poor probe quality, given their RNA-seq read numbers not being low. Resembling Ham protein expression (Fig. 1G), ham mRNAs were detected in the TE terminal epithelium (Fig. 6A), confirming the specificity of the SCRINSHOT method. Similarly, wnt2, shg and shot all exhibited high expression in the TE terminal epithelium (Fig. 6A).

Figure 6.

In total, 25 genes exhibited detectable SCRINSHOT signal in the wt SV and AG, while 29 displayed signals in the wt TE (Supplementary Table 3). Within the TE, 17 out of 29 were specifically localized in the TE terminal epithelial cells, with the remaining 12 detected only in other TE cells (Fig. 6, Supplementary Fig. 7, Supplementary Table 3).

Subsequently, our analysis focused on the genes expressed in the TE terminal and SV epithelia. While 17 genes exhibited expression in both epithelial ends (Fig. 6A-B), 8 were exclusively expressed in SV epithelial cells (Fig. 6F, Supplementary Table 3), underscoring the heterotypic nature of the TE and SV epitheliums. Notably, certain genes (CG11406, AP-2alpha, GstE11, Tl, spz5 and spz6) demonstrated temporal dynamic expression in the TE epithelium: spz5 and CG11406 are present during the 20-25h APF timeframe, prior to fusion, but gone during the 28-32h APF period when the TE and SV are undergoing fusion; AP-2alpha, GstE11, Tl and spz6 showed the opposite trend, being expressed only later (Supplementary Fig. 7B).

In the GD, while the majority of genes exhibited ubiquitous expression, wnt2 and spz were predominantly expressed in the leading cells within the SV, contrasting with their expression in the remaining GD (Fig. 6B). Additionally, the expression patterns of fz2, fz3 and Tl displayed spatial specificity: fz2 exhibited higher expression levels in the SV compared to the AG, fz3 demonstrated the opposite pattern with higher expression in the AG, and Tl was primarily expressed in the myoblasts (Fig. 6B). These findings imply that Wnt2 and Tl signaling may have cell-type-specific roles in the fusion process. Thus, SCRINSHOT mapped 30 candidate genes in the developing GD and TE, with 29 of them display spatial or temporal dynamics.

Wnt2 and Tl signaling pathways are key mediators of Ham function in RS tissue assembly

Next, we aimed to identify genes exhibiting altered expression in ham mutant SV and TE. As hamSK1 mutant TE lacks terminal epithelial cells, we excluded TE samples and only quantified SCRINSHOT signal levels in the developing SV between wt and the hamSK1 mutant. In contrast, ham RNAi knockdown resulted in fusion defects but only mildly impacted SV morphology and TE terminal epithelial cell number (Fig. 2G-H). These samples allow to assess gene expression alterations in both the SV and TE epithelia.

ham fluorescent intensity appeared comparable between wt and mutant samples (Fig. 6E, Supplementary Fig. 7E), suggesting that ham mutant RNAs, containing the CRISPR-generated indels, remain stable at this stage. In the hamSK1 mutant SV, seven genes became downregulated, while two exhibited upregulation. The downregulated genes were four Wnt2 pathway genes (wnt2, fz, fz2, fz4), one Tl pathway gene (spz), and shg (Fig. 6B-C). Notably, wnt2, fz, spz and shg also displayed downregulation in ham RNAi SV (Fig. 6D-E), suggesting that these genes and their associated signaling pathways may be directly involved in the fusion process downstream of Ham. The upregulated genes in hamSK1 mutant SV include GstE11 and CG11069 (Supplementary Fig.7C, E), which exhibited low expression in the leading cells of wt SV epithelium, but increased expression under the hamSK1 condition. In ham RNAi knockdown animals, CG11069 was again upregulated, while GstE11 did not change expression. Instead, another ligand gene in Tl signaling (spz5) displayed increased expression (Supplementary Fig.7D, F). On the TE side, only four genes showed expression alteration in ham RNAi animals, including ham itself, shg, fz and fz2 Supplementary Fig.7G), implicating functional requirement of these genes in both TE and SV epithelia.

The observation that knocking down Tl resulted in decreased male fertility (Fig. 5B, Supplementary Table 2), coupled with the expression of multiple spz ligand genes in the developing TE and SV, suggests that the Tl signaling pathway might be required for assembling the TE and SV. To investigate this possibility, we examined male flies with spz knockdown driven by hamRSGal4. These animals produced a comparable number of progenies to control animals. The lack of phenotype in this condition could be attributed to inefficient knockdown or genetic compensation by another spz gene, like spz5 whose expression increased in ham RNAi animals when spz expression was reduced. In addition to Tl, two other Tl related genes (Toll-7 and Toll-9) exhibited moderate number of reads in the RNA-seq data. We then tested male fertility in Toll-7 and Toll-9 knockdown animals. Toll-9 knockdown led to a drastic reduction in progeny numbers compared to the RNAi control (Supplementary Fig. 8A). Consistent with the sterility, the adult SV in the Toll-9 knockdown male lacked sperms (Supplementary Fig. 8A), suggesting failed connection between the TE and SV. Together, these results pinpoint an essential role of Tl signaling in male RS development.

To validate the role of Wnt2 signaling in RS development, we investigated four genetic conditions, aimed at modulating its level and activity: wnt2^O (a wnt2 null allele), wnt2 overexpression driven by hamRSGal4, fz and fz2 RNAi knockdown driven by hamRSGal4. The wnt2^Omutant TE and GD never connect, suggesting that Wnt2 is essential for the fusion process. The wnt2^O mutant TE lacks the pigment cell and myocyte layers (Fig. 7A), consistent with its reported function in these two cell types [17, 39]. The wnt2^O mutant GD failed to form elongated SV and AG, suggesting that Wnt2 is required for GD patterning and differentiation. Moreover, overexpression of wnt2 also resulted in fusion defects and mis-localization of TE epithelial cells (Fig. 7B), a phenotype reminiscent of shg knockdown animals. Indeed, wnt2-overexpressing cells displayed reduced levels of E-Cad, suggesting that Wnt2 signaling might balance E-Cad levels to maintain TE epithelial cell properties. Consistently, fz2 knockdown animals exhibited comparable phenotypes, including mobilized TE epithelial cells and failed tissue fusion (Fig. 7C). fz2 and fz knockdown also decreased male fertility (Fig. 7D). Collectively, these results suggest the importance of optimal Wnt2 signaling levels in male RS development and epithelial tissue fusion.

Figure 7.

Discussion

In this study we demonstrate that the transcription factor Ham plays a crucial role in controlling epithelial fusion during Drosophila RS development. Early in embryonic development, Ham determines the cell fate of msSGPs, which are progenitor cells for TE terminal epithelial cells. Later on, Ham directly regulates the differentiation of epithelial cells in both the TE and SV by orchestrating a gene expression program in epithelial cells. While our primary focus is on male RS, we found that Ham is also expressed in cells at the fusion site of the oviduct and ovary. Mutant females lacking functional Ham exhibit a non-functional oviduct epithelial tube, suggesting that Ham likely employs a similar mechanism to control epithelium fusion in females.

By examining ham mutant phenotypes and its downstream target genes, we identified several classes of novel factors involved in fusion process of SV and TE. These genes, directly or indirectly controlled by Ham, encompass a diverse array of molecular functions. This discovery provides a comprehensive gene network for future investigation of epithelial fusion processes. The cell-type-specific and tissue-specific spatial distribution of these genes suggest that they may regulate various aspects of fusion, including cell migration, intercellular communication between myoblasts and epithelial cells, specification of the leading cells, and other steps in the fusion process. For instance, Tl and its related receptors is known to mediate cell adhesions [55]. We can speculate that Tl signaling and the molecule itself promote interaction between myoblasts and myoblast and epithelial cells (Fig. 7E), based on high expression of Tl in myoblasts and presence of multiple Spz ligands in the TE and SV epithelial cells. Moreover, Wnt2, fz and fz2 are highly expressed in the leading cells of TE and SV epithelia, and also required for TE and SV connection, suggesting that Wnt2 signaling mediates epithelial cell communication, remodeling and fusion.

Despite being discovered more than twenty years ago [28], research on the function of Ham function has primarily focused on its roles in the nervous system. Our findings highlight a crucial function of Ham in epithelial tissues. Apart from the RS, we observed high expression of the Ham protein in other epithelial tissues including the Malpighian tubes (akin to the kidney in insects) and the intestine (data not shown). Furthermore, orthologs of Ham, PRDM16 and PRDM3/MECOM, exhibit high expression in epithelial tubular structures, such as Lung, pancreas, kidney, urinary bladder and gastrointestinal tract and lung, according to observations in the Human Protein Atlas (Supplementary Fig. 9. https://www.proteinatlas.org/ENSG00000142611-PRDM16/single+cell+type; https://www.proteinatlas.org/ENSG00000085276-MECOM/single+cell+type) [56]. Notably, Seminal vesicle and ductus deferens, the equivalent structure to fly seminal vesicle, are the top one and three tissues with the highest PRDM16 expression (Supplementary Fig. 9A). This suggests a potential role for these proteins in controlling epithelial integrity and fusion in mammalian organogenesis. Indeed, we observed that the pancreas duct volume in mouse embryos lacking Prdm16 is significantly reduced (data not shown). Although the exact cellular changes and the underlying molecular mechanisms in this context await further investigation, this phenotype underscores a crucial function of Prdm16 in pancreatic epithelial duct formation.

We show that compared to Ham function in the nervous system, the PR-containing isoforms are indispensable for RS epithelial fusion and differentiation, and that Ham can directly activate epithelial gene expression. It remains to be determined whether Ham has distinct molecular functions or it regulates a subset of common molecular pathways in these two contexts. Dendrite morphogenesis involves outgrowth and branching, somewhat analogous to epithelial tube growth and branching. For instance, molecular mechanisms regulating polarized protein trafficking are crucial for both dendrite and epithelial tube branching [57]. Additionally, several transcription factors like Sequoia [58, 59], Tramtrack [60, 61] and Cut [62, 63] play roles in both dendrite morphogenesis and tracheal tube fusion. These examples suggest that there might be more unrecognized common mechanisms between dendrite and epithelial morphogenesis. It is of interest to test whether the effectors we identified in the RS epithelium are involved in dendrite morphogenesis in the nervous system.

We observed a unique aspect of epithelial fusion in the Drosophila RS. Unlike many epithelial fusion processes where two homologous ends merge, such as those in mammalian neural tube closure and Drosophila tracheal formation etc [1, 8, 10], the two epithelial ends in the Drosophila RS are heterotypic. Specifically, the TE terminal epithelial cells express higher levels of E-Cad and lower levels of Cor compared to SV epithelial cells. In response to the reduction of E-Cad or ectopic expression of wnt2, TE terminal epithelial cells become motile, resembling a process akin to epithelial-to-mesenchymal transition. This suggests that TE terminal epithelial cells may be more invasive during epithelial fusion. On the other hand, SV epithelial cells, despite originating from mesenchymal cells, exhibit typical epithelial cell characteristics, including regular Cor and E-Cad expression. Notably, many genes are expressed exclusively in the SV epithelium, distinguishing it from the TE epithelium. Heterotypic epithelial fusion occurs in many developmental processes, such as digestive system formation, vertebrate optic fissure closure, and mammalian urogenital development. Failure of epithelial fusion can lead to a wide variety of developmental diseases such as oesophageal atresia, spina bifida, ocular coloboma, congenital abnormalities of the kidney, urinary tract and persistent cloaca. Studying the basic processes and key players of normal epithelial tissue fusion in an in vivo context will help to understand the pathogenesis of these diseases.

Acknowledgements

We thank the imaging facility (IFSU) of Stockholm University, the BEA sequencing facility at Karolinska Insitute, and the National Genomic Institute of Scilife Laboratories, Sweden, for providing service and support. The Bloomington Stock Center and the Drosophila community provided important reagents and fly stocks. We thank Christos Samakovlis’s lab for helping to set up SCRINSHOT and Eric Lai’s lab for the initial Ham reagents. Funding from Australian Research Council Centre of Excellence for the Mathematical Analysis of Cellular Systems (CE230100001) and ANU Future Scheme to J.W. The project was supported by project grant from Swedish Research Council (Vetenskapsrådet, 2020-03543) and Swedish Cancerfonden (CAN 22 2161 Pj) to Q.D.

Materials and methods

Fly strain culturing and generation of transgenes

All fly stocks were maintained at 22°C. The ham^PRΔ mutants were created using CRISPR: transgenic flies carrying single guide RNA targeting the PR domain containing isoforms were crossed into the vas-Cas9 transgenic flies. Frame-shift mutations were identified by PCR and Sanger sequencing. Transgenic flies were created at BestGene, Inc.

The ham Gal4 lines and RNAi stocks were obtained from Bloomington Stock Center. Knocking down experiments were performed at 25 °C and with a regular 12-hr light–dark cycle.

To generate ham null mutant somatic clones in the TE terminal epithelium, hamsk1 FRT40A/CTG males were crossed with virgin females carrying the genotype hsFLP(X); hsGFP.nls FRT40A. Embryos aged 0-5 hours were collected and developed at 25°C until the 2nd instar larval stage. GFP-negative larvae, indicating the genotype hsFLP(X); hsGFP.nls/hamsk1 FRT40A, were then selected. Head shock treatments at 37°C for 1 hour were administered three times during the 2nd and 3rd instar larval stages to induce somatic clones. Prior to dissection, another 1-hour head shock at 37°C was performed to induce GFP expression, marking the wild-type cells.

Fertility test

For ham mutants, single male or female with indicated genotype was mated with two w¹¹¹⁸ virgin females or two w¹¹¹⁸ males on day 0 of the experiment. Crosses were flipped every 2 days for 3 consecutive times. Progenies from the second vial with eggs laid between day 2-4 were counted after 2 weeks.

For the RNAi screen, individual UAS-RNAi males were crossed with females carrying UAS-dcr2 and hamRSGal4. The resultant F1 male progenies (2-5 days old) with the genotype UAS-dcr2; hamRSGal4>UAS-RNAi were crossed to w¹¹¹⁸ virgin females (1-2 days old) on day 0 of the experiment. In each test, four males were crossed with three females. Crosses were flipped on day2 and day3 for pre-laying. After flipping on day3, crosses were kept in the new vial for egg laying for a specific time period (GroupA: 48hours; GroupB: 16-18h; GroupC: 7-8h). If at least three males and two females were alive, then the flies were transferred to a new vial, the old vial was kept, and all progenies were counted.

For fertility test of the Wnt2 and Tl signaling pathways components, UAS-RNAi males were crossed to virgin females of the genotype UAS-dcr2; hamRSGal4. Two F1 male progenies (1-2 days old) with the genotype UAS-dcr2; hamRSGal4>UAS.RNAi were crossed to three w¹¹¹⁸ virgin females (1-2 days old) on day 0 of the experiment. Crosses were flipped every day for 4 consecutive times. Progenies from the forth vial with eggs laid between day 3-4 were counted after 2 weeks. Crosses were flipped every day for 4 consecutive times. Progenies from the forth vial with eggs laid between day 3-4 were counted after 2 weeks.

Immunostaining

For embryo staining: embryos were collected, dechorionated with bleach and fixed with 4% formaldehyde in PEM (0.1M PIPES, 2mM EGTA, 1 mM MgSO4, pH 7) for 20 min. For immunostaining, embryos were rehydrated, washed with PBT (0.3% Triton X-100 in PBS) and blocked with PBSBT (0.3% Triton X-100, 1% BSA in PBS).

For tissue staining: tissues were dissected from stage defined larvae or pupae in 1xPBS within 30 min and fixed in 4% paraformaldehyde in PBS for 20min. For immunostaining, dissected tissues were washed with PBT (0.1% Triton X-100 in PBS) and blocked with PBSBT (0.1% Triton X-100, 1% BSA in PBS).

For both embryo and tissue staining, the primary antibodies were diluted with PBSBT (0.1% Triton X-100, 1% BSA in PBS) and incubated at 4°C over night. After extensive washing with PBSBT (15min x 4 times), the secondary antibodies were added in 1:500 dilution and incubated at room temperature for 2h. After washing with PBT (0.1% Triton X-100 in PBS), the samples were counter-stained with DAPI and mounted in Vectashield (Vector Laboratories).

The following primary antibodies were used: Rabbit anti-Ham (1:300) (Gift from Eric Lai’s lab), Ms anti-V5 (1:500, R960-25, invitrogen), chk anti-GFP (1:2000, ab13970, Abcam), Rat anti-vasa (1:50, anti-vasa-c, Developmental Studies Hybridoma Bank, DSHB), Ms anti-ABD-B (1:50, 1A2E9-c), Ms anti-SXL (1:50, M18-c), rat anti-DE-Cad (1:50, DCAD2, DSHB), mouse anti-Crb (1:10, Cq4 DSHB), mouse anti-Coracle (1:100, C615.16 DSHB). Secondary antibodies conjugated to Cy3 or Cy5 or Alexa Fluor-488 (Jackson ImmunoResearch) were used.

Quantification of signal intensity from fluorescent immunostaining

For E-Cad and Crb in Figure 2, “raw intensity density” from manually defined regions (using free-hand drawing) along the apical side of the SV and TE terminal epitheliums was obtained in ImageJ. For each sample, 2 regions in the SV and 3 regions in the TE terminal epithelium were averaged. The regions used to determine relative E-Cad fluorescence in hamSK1 TE somatic clones were at the apical side between cell junction. The length of Cor-stained cell junction was measured using ImageJ. A straight line was drawn over the Cor staining with the ‘Freehand Line tool’. For each sample, the average length from 10 cell-cell junctions were used. In all experiments, images for the genotypes in comparison were always processed using exact settings. At least three biological replicates were included in statistical analysis.

Western Blot

Embryos from w1118, ham^PRΔ/CTG, ham^SK1/CTG were collected and aged to embryonic stage S13 to S16. In each blot, 50-100st GFP+ or GFP-embryos were used. The selected embryos were homogenized directly in 1xSDS loading buffer (62.5 mM Tris-HCl pH 6.8, 2.5 % SDS, 0.002 % Bromophenol Blue, 0.7135 M (5%) β-mercaptoethanol, 10 % glycerol) and denatured at 98 °C heat blot for 5 min. After centrifuged, supernatant was taken for Western-Blotting with 8% of SDS-PAGE. Rabbit anti-Ham (1:2000) primary antibody was used in blotting. To minimize unspecific signal from IgG heavy chains, light chain specific secondary antibody of mouse anti-Rabbit IgG (1:40000 Jackson ImmunoResearch) was used, and the signals were developed with ECL Plus reagent (GE Healthcare).

Cell culture and Luciferase assays

To generate different variants of Ham, the open reading frames of HamFL, Ham-NZF-vp16, Ham-CZF-vp16, and ham^PRΔ were PCR amplified and cloned into the pAC_V5 (nitrogen) vector. The enhancer fragments were cloned in to the p2xTetO-Firefly luciferase vector. All transfections were performed using Drosophila S2-R+ cells grown in Schneider Drosophila medium containing 10% fetal calf serum. Cells were co-transfected with the Ham expression construct, 2xTetO-Firefly luciferase and pAc-Renilla plasmids in 96-well plate using the Effectene Transfection kit (Qiagene). Luciferase assays were performed and measured using the Dual Luciferase Assay System (Promega). Expression was calculated as the ratio between the firefly and Renilla luciferase activities. Three rounds of transfections were performed on different days and different cell populations and considered as three biological replicates. Within each biological replicate, 4 wells of cells were transfected with the same DNA mixture and measured and thus considered as 4 technique replicates. In total, for each DNA combination, the values from the 12 wells of transfection were calculated, averaged and plotted in Fig. 5D-F. Cloning oligos are provided in Supplementary table 4.

SCRINSHOT

The Drosophila reproductive system was dissected at defined ages in 1xPBS and fixed in 4% paraformaldehyde in PBS for 20min. After washing with PBST 0.1% (Triton X-100) for 2 x 15 min, fine dissected samples were placed on Poly-lysine coverslip and freezed at -80°C after drying on a heater.

The SCRINSHOT experiments were carried out according to the published method [54]. In brief, 3 padlock probes and 3 corresponding detection probes were designed for each gene of interest. The slides with fine dissected samples were pretreated at 45°°C to reduce moisture and fixed in 4% PFA in 1X PBS, followed by washing in PBS Tween-20 0.05% twice. Samples were then dehydrated with 70%, 85%, 100%ETOH and air dry. The SecureSeal hybridization chamber (GRACE BIO-LABS, 621501) was then mounted to cover samples on the slide. Permeabilization of tissues were first done by treating with 10µg/mL Proteinase K (Invitrogen) in PBSTween-20 0.05% for 7 minutes followed by washing with Glycine (2mg/mL in PBS Tween-20 0.05%) once and with PBS Tween-20 0.05% for 3 x 5 minutes. After this, samples were further Permeabilized in 0.1M HCl for 3min, followed by washing in PBS Tween-20 0.05% twice. Samples were then blocked in a probe-free hybridization reaction mixture of 1XAmplifase buffer (Lucigen, A1905B), 0.05M KCl, 20% Formamide deionized (Millipore S4117), 0.1uM Oligo-dT, 0.1ug/ul BSA (New England Biolabs, B9000S), 1U/ul RiboLock (Thermo, EO0384), and 0.2ug/ul tRNAs (Ambion, AM7119). Then hybridization of padlock probes was done by incubating samples with padlock probes (with the concentration of each one 0.01uM) mixed in blocking reagents used before (no oligo dT used in this step). The slide was then put into a PCR machine to denature at 55°C for 15 minutes and hybridize at 45°C for 120 minutes. Padlock probes were then ligated by using SplintR ligase (NEB M0375) at 25°C for 16 hours followed by RCA (rolling cycle amplification) at 30°C for 16 hours by using phi29 polymerase (Lucigen, 30221–2) and RCA primer1. Then a fixation step was applied to stabilize RCA product in 4% PFA for 15 minutes followed by washing in PBS Tween-20 0.05%. Then the hybridization of the first 3 genes was done by mixing all 3 3’ fluorophore-conjugated detection probes of each gene in reaction reagent (2XSSC, 20% Formamide deionized, 0.1ug/ul BSA, and 0.5 ng/μl DAPI (Biolegend, 422801) followed by hybridization at 30°C for 1 hour. The slides were washed in 20% formamide in 2X SSC and then in 6X SSC, followed by dehydration in 70%, 85%, 100% Ethanol until the chamber was removed. Then the samples were preserved in SlowFade Gold Antifade mountnant (Thermo, S36936) and kept in dark before imaging. After image acquisition, the first 3 detection probes were removed by using Uracil-DNA Glycosylase (Thermo, EN0362), and the slides were ready for the next round of detection probe hybridization. The procedure was repeated until all genes were hybridized and imaged. Images were acquired with an airyscan-equipped confocal microscope system (Zeiss LSM 800, Carl Zeiss), and ZEN 2.5 Blue edition software was used for image processing.

Quantification of SCRINSHOT signals

Similar to quantification of immunofluorescent signal, fluorescence intensity of SCRINSHOT for each gene was measured in defined areas within images of the SV, EjD, and TE in 2D projections. Squares were drawn to define areas at the leading side of the SV, the middle region of the Ejd and the TE terminal. The total fluorescence intensity was first normalized to area size. The resultant value for ham mutant or RNAi samples was then contrasted to the normalized intensity of control samples.

CUT&TAG

CUT&TAG was performed according to the published method [37]. In brief, TE and grossly tissue of the posterior abdomen were dissected from 18-25h APF w1118 and pBAC_ham_sfGFP animals in PBS within 30 min. For grossly tissues, the dissected samples were homogenized in 2ml NE1 buffer followed by incubation on ice for 10 min. After filtering out the grossly cuticles, nuclei were collected by spinning at 1300xg and resuspended in 1ml PBS. For TE, nuclei extraction was done by simply pipetted samples in 200ul NE1 up and down for 20-30 times and incubated on ice for 10 min. Light fixation was performed with 0.1% formaldehyde at RT for 2 minutes and neutralized with 75mM Glycine. Then the nuclei were washed by PBS and wash buffer 1 time each and kept in 200ul wash buffer on ice until enough samples were collected (30-50st animals per/antibody). Resuspended nuclei in 1.2ml wash buffer and mixed with 20ul Concanavalin A-coated magnetic beads (Polyscience 86057) with soft vortexing. Aliquoted samples into two parts immediately and incubated for 10 min with rotation to allow the binding of the nuclei with the beads. The nuclei/beads mix was then blocked with 1mL cold antibody buffer for 10 min and resuspended in 150ul cold antibody buffer. For w1118 samples, 1uL Rabbit anti-Ham or 1ul Rabbit IgG (SIGMA-ALDRICH, I5006) was added for an overnight incubation at 4°C. For pBAC_ham_sfGFP samples, 1ul Rabbit anti-GFP (G10362, invitrogen) or Rabbit IgG was used. The beads/nuclei/antibody mix was then washed with Dig-wash buffer and incubated with secondary antibody (1:100 dilution) for 1 hour. After further washes with the Dig-wash buffer, a pA-Tn5 adaptor complex in Dig-300 buffer was added to the beads for 1 hour reaction. The beads were then washed with Dig-300 buffer before the incubation with 300ul Tagmentation buffer at 37°C for 1 hour. Then 10ul 0.5M EDTA, 3uL 10% SDS and 2.5uL 20 mg/mL proteinase K (Invitrogen) were used to stop the reaction. Then the resultant DNA was purified with DNA clean & Concentrator kit (ZYMO Research D4013) and eluted in 25uL Elution buffer. To generate libraries, 21uL fragment DNA was mixed with 2uL 10uM Universal i5 primer, 2uL 10uM uniquely barcoded i7 primer and 25uL PCR master mix (NEB Phusion® High-Fidelity PCR Master Mix with HF Buffer, M0531). The PCR condition was as follows: 72°C 5min, 98°C 30s, repeat 12 times (98°C 10s, 63°C 10s), 72°C 1 min and hold 4°C. The libraries were cleaned with standard Ampure XP beads as previously described. Libraries from four biological replicates were produced for each condition.

We mapped the CUT&TAG samples of w1118 and pBAC_ham_sfGFP in TE and TG tissues to the dm6 genome assembly using Bowtie2 with the following parameters: bowtie2 --end-to-end --very-sensitive --no-mixed --no-discordant --phred33 -I 10 -X 700. The coverage for these samples was generated using bedtools genomeCoverageBed and normalized to library depth to yield reads per million per base (RPM). For peak calling, we employed SEACR [64], which is tailored for sparse chromatin profiling data like CUT&TAG. We first used a relaxed FDR cutoff of < 0.5 for each replicate (SEACR_1.3.sh normalized_coverage 0.5 norm stringent output_peaks), we then identified common peaks across replicates for each condition. To select confident peaks for w1118 and pBAC_ham_sfGFP, we overlapped their peaks within TE and TG separately, choosing the top 1800 peaks from TE and the top 6000 from TG. Peak annotation was performed by HOMER’s annotatePeaks.pl. Peak overlap analysis was performed using HOMER’s mergePeaks function with the default -d given parameter, ensuring that only directly overlapping peaks were merged. This method preserved the integrity of the original peak boundaries and maintained the specified distances between peaks from the initial peak calls. De novo motif discovery for peaks in up-regulated or down-regulated genes, positioned 2000 bp upstream and 500 bp downstream of the TSS in TE and TG tissues, was performed using the MEME-ChIP software. The parameters used were: -ccut 100, -meme-p 5, -dna, -meme-mod anr, -minw 5, -maxw 15, and -filter-thresh 0.05.

RNA-seq

Total RNA was extracted from finely dissected genital discs or testis samples from 18-25h APF animals using Trizol reagent (Invitrogen). RNA quality was tested by Agilent Bioanalyzer.

RNA-seq libraries were made using the Illumina Truseq Total RNA library Prep Kit LT. Sequencing was performed on the Illumina Hiseq2500 platform. Following the trimming of adapter sequences with Trimmomatic, RNA-seq reads from replicated wild type (x3) and mutant samples (x3) were aligned to the Drosophila melanogaster (dm6) genome assembly utilizing HISAT2. Gene annotations were sourced from UCSC. The DESeq2 Bioconductor package was employed to identify differentially expressed mRNAs between w1118 and the ham^PRΔ/ham^PRΔ mutant, as well as between +/Df and ham^SK1/Df mutant, with a subsequent estimation of the false discovery rate (FDR). Genes with an FDR < 10% and absolute Log2FC>0 were considered differentially expressed. Details of the calculated differential expression are provided in Supplementary table 5.

STRING gene network analyses

Ham activated (genes downregulated in hamSK1 and bound by Ham in TG) and repressed (genes upregulated in hamSK1 and bound by Ham in TG) gene networks are constructed using String, MCL clustering (find natural clusters the stochastic flow, inflation parameter, 2. Disconnected genes are hidden. The position of the clusters is manually adjusted to fit related GO categories. The links to the gene networks in STRING: https://string-db.org/cgi/network?taskId=bbF8IgMTqf1M&sessionId=bzDIAZvMgWsi https://string-db.org/cgi/network?taskId=bhKCcMoLlbwh&sessionId=bzDIAZvMgWsi

Additional information

Author contributions

Conceptualization and project design by QD, experiments conduction by HW, preliminary test in pancreas by IA and LB, computation analysis by JW. Data interpretation by HW, JW and QD. Manuscript writing by QD with input from HW and JW.

Declaration of the use of AI

ChatGPT was used to improve readability of the text, and the content was thoroughly checked and edited by the authors to ensure accuracy.