A miRNA screen identifies GSC fate regulators

To explore the potential role of miRNAs in GSC fate, we obtained a collection of 90 miRNA mutant lines covering 111 Drosophila miRNAs from the Bloomington Drosophila Stock Center (S1 Table) [45]. We examined the ovaries of a collection of homozygous viable mutations affecting 54 miRNAs. An antibody to Hts (Hu li tai shao) labels the spectrosome organelle of GSCs and CBs and its branched derivative, the fusome, which connects sister cells of differentiating cysts derived from CBs. GSCs are evident from their anterior location and the presence of a spectrosome directly adjacent to a Cap cell. Significantly fewer GSCs were observed for the loss of function of miR-971, miR-975-976-977, miR-11, miR-219, or miR-999 (S1 Fig), suggesting a role in GSC maintenance. Defects in GSC or initial CB differentiation can be recognized by the accumulation of spectrosome-containing cells. Similar to previous findings, control germaria usually contained 2 or 3 GSCs and 1 CB to give an average of 3.36 (n = 81) spectrosome-containing cells per germarium (Figs 1B and S1A and S2 Table). We found that germaria with a loss of function of miR-970, miR-124, miR-986, or miR-957 showed an accumulation of undifferentiated germ cells with a spectrosome posterior to GSCs (Fig 1C and S2 Table). We measured the prevalence of the phenotype by scoring the percentage of germaria with 5 or more spectrosome-containing cells, and we referred to the additional cells posterior to normal GSC locations as CB-like cells. We found that approximately 50% (n = 63) of miR-970, 70% (n = 70) of miR-124, 65% (n = 52) of miR-986, and 83% (n = 60) of miR-957 mutant germaria contained extra CB-like cells, while the incidence in the controls was less than 10% (Fig 1D). Of these miRNAs, miR-124 was of special interest because it is conserved among vertebrates and invertebrates [46–48] and has a number of known functions, including roles in Drosophila neuroblast proliferation, courtship, circadian locomotor, and rhythmic behaviors [49–52], as well as the inhibition of differentiation of mouse bone marrow mesenchymal stem cells and pre-cartilaginous stem cells [53,54].

miR-124 expression in TF and Cap cells is critical for GSC daughter cell differentiation

We tested another 2 independent miR-124 null allele lines (miR-124[6] and miR-124[7]) [49] and found similar phenotypes, with approximately 70% of germaria containing 5 or more cells with spectrosomes in each case (Fig 1E–1G); the strength of this phenotype is similar to several prior studies of genotypes with impaired germline differentiation [28,41,55]. We also tested fertility. Both the number of eggs laid per day and the frequency of larval hatching per egg were significantly reduced in miR-124 mutants relative to the w¹¹¹⁸ control (S2 Fig). To investigate the expression pattern of miR-124, we used a reporter line (miR-124::dsRed) for which RFP expression controlled by the promoter of miR-124 was previously found to mirror endogenous miR-124 expression in embryos [49]. We detected specific miR-124 expression in the anterior-most part of the germarium, and also in early follicle cells and some FSCs in region 2b but very little expression in between (Fig 1H and 1H’; quantitation in S3 Fig). The miR-124-positive cells in the anterior of the germarium mostly overlapped 2 cell types, TF and Cap cells, which both express the marker Lamin C (LamC) (yellow arrowheads in Fig 1H and 1H’) [27,56]. There was also occasional expression in the most anterior ECs, immediately posterior to Cap cells (white arrowhead in Fig 1H and 1H’).

To investigate whether miR-124 function is important in TF and Cap cells, we attempted to rescue the miR-124 loss-of-function differentiation defect through spatially selective expression of miR-124. We expressed UAS-miR-124 specifically in TF and Cap cells using a bab1-GAL4 line [57]. We first confirmed that expression of this bab1-GAL4 line (bab^Agal4-5) is limited to TF and Cap cells and does not extend to ECs in adult ovaries (S4A and S4A’ Fig). A different bab1-GAL4 line, employed in some other studies, has a similar pattern but also has detectable expression in ECs [58]. The bab1-GAL4 line we used is highly specific in adults but it is expressed in a broader group of somatic cells of the developing female gonad [57]. We therefore used a temperature-sensitive conditional GAL4/GAL80^ts system to drive UAS-miR-124 expression only in adults (S5A–S5B’ Fig). Flies of appropriate genotypes were raised at 18°C to suppress GAL4 activity during developmental stages. After eclosion, half of the adults were switched to 29°C to inactivate GAL80^ts and thereby permit GAL4 activation of UAS-miR-124, while the remaining half were kept at 18°C. Germaria were examined 5 and 10 d later. The germaria of miR-124 mutant animals maintained at 18°C throughout (miR-124[6], GAL80^ts; bab1>miR-124 and miR-124[7], GAL80^ts; bab1>miR-124) contained extra CB-like cells at high penetrance (Fig 2A–2C), consistent with the absence of any potential rescue activity from the UAS-miR-124 transgene. In contrast, when incubated at 29°C for 5 d or 10 d as adults, most germaria contained the normal number of spectrosome-containing cells, indicating full phenotypic rescue (Figs 1D and 2A–2C). We also tested the converse temperature shift, raising flies at 29°C and then moving half to 18°C after eclosion. Flies raised and maintained at 29°C throughout exhibited only a low basal frequency of excess CB-like cells (<20%) but this phenotype was increased to a penetrance of more than 65% by 5 d and more than 75% by 10 d at 18°C for both miR-124[6] and miR-124[7] (S5C Fig). We conclude that miR-124 function is required in adult TF or Cap cells to promote GSC progeny differentiation.

We also performed the same type of experiment with C587-GAL4 [23,26,27], which is specifically expressed in ECs but not TF or Cap cells (S4B and S4B’ Fig). As expected, the germaria of animals maintained at 18°C throughout (C587; miR-124[6], GAL80^ts>miR-124 and C587; miR-124[7], GAL80^ts>miR-124) contained extra CB-like cells at high penetrance (Fig 2D). When incubated at 29°C for 5 d or 10 d as adults, the germaria still contained an excess of spectrosome-containing cells (Fig 2D), albeit marginally fewer than at 18°C. These results are consistent with miR-124 acting in TF or Cap cells, where miR-124 is strongly expressed, to promote GSC differentiation. It is possible that miR-124 activity in anterior ECs can also sometimes contribute to GSC differentiation, reflecting occasional expression in those cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. miR-124 in GSC niche cells affects daughter cell differentiation.

(A and B) Ectopic expression of UAS-miR-124 in the TF and Cap cells of miR-124[6] (A) and miR-124[7] (B) mutant backgrounds using bab1-GAL4. Germaria were immunostained for Hts (red), Vasa (green), and DAPI (blue). In (A) and (B), the GSCs are highlighted by dashed circles. Scale bar: 10 μm. (C and D) Animals carrying UAS-miR-124 and GAL80^ts; bab1-GAL4 (C) or UAS-miR-124 and C587-GAL4; GAL80^ts (D) were raised at 18°C up to eclosion and then maintained at 18°C or 29°C for the number of days indicated before ovary dissection. The percentage of germaria carrying 5 or more spectrosome-containing cells is shown, and the number of analyzed germaria is above each bar. Significance of 18°C versus 29°C for the same time period was determined by Fisher’s exact two-sided test (* P < 0.05; *** P < 0.001; ns: not significant). (E–G’) Germaria from wild-type females (E–E’), miR-124[6] (F–F’), and miR-124[7] (G–G’) mutants were immunostained for a marker (pMad) of BMP pathway activity. pMad (red) in germline cells (marked by Vasa antibody, green) was normally present only in GSCs (E–E’, highlighted by dashed circle) but was present in additional cells (indicated by white arrowheads) in the germaria of miR-124 mutants (F–F’, miR-124[6]; G–G’, miR-124[7]). (E), (F), and (G) show the merging of the 3 channels of pMad, Vasa, and DAPI (blue); (E’), (F’), and (G’) show pMad stained images in black and white. Scale bar: 10 μm. (H) Average numbers of cells with pMad staining per germarium in wild-type and 2 miR-124 mutants. Left to right: n = 103, 83, 113 biologically independent germaria. Data are presented as the mean ± SEM. Significance was analyzed by Kruskal–Wallis one-way ANOVA with Dunn’s test (*** P < 0.001). The raw data underlying panels C, D, and H are available in S1 Data. BMP, bone morphogenetic protein; GSC, germline stem cell; TF, terminal filament.

https://doi.org/10.1371/journal.pbio.3002515.g002

miR-124 limits BMP signaling in the CB differentiation niche

BMP signaling elevation is responsible for most or all CB differentiation defects caused by a variety of genetic alterations in ECs [26,55,59,60]. We examined BMP signaling in miR-124 mutant germaria using phosphorylated Mad (pMad) as an indicator of BMP signaling activity. In GSCs, BMP signaling leads to the production of a strong pMad signal, whereas CBs and mitotic cysts show very little pMad signal [61]. In control germaria, pMad signal clearly above background levels was restricted to GSCs, with an average of 1.9 positive cells per germarium (Fig 2E, 2E’, and 2H). In contrast, mir-124[6] and mir-124[7] mutant germaria included ectopic pMad staining in cells that were not adjacent to Cap cells, in addition to normal GSC staining (Fig 2F–2G’ and quantified in S6 Fig). The average total number of pMad-positive cells was 3.5 (n = 83) for mir-124[6] and 3.7 (n = 113) for mir-124[7] (Fig 2H). The number of extra pMad-positive cells was slightly smaller than the increase of spectrosome-containing cells in both mir-124[6] and mir-124[7] germaria. Similar comparative results have been reported in almost all studies of germline differentiation defects and likely simply represent the different sensitivities of the 2 assays [23,27,62]. Accordingly, even among GSCs in control germaria, some do not have strong pMad staining. The total number of extra cells with a spectrosome or pMad staining was lower in miR-124 mutant germaria than in several other situations where EC function was genetically compromised [27,63,64], and the extra cells were always observed close to the normal location of CBs. These observations suggest that the key deficit in the somatic cell environment of germline cells is confined to the anterior of the germarium and that differentiation is mostly delayed rather than arrested. This, in turn, suggests that the function of only the more anterior ECs is compromised by the absence of miR-124 in TF and Cap cells.

EC cell bodies reside next to the basement membrane around the circumference of the germarium, with cellular processes reaching into the interior to wrap developing germline cells. The presence of these long cellular processes is essential for germ cell differentiation [26,55,65]. Some perturbations in ECs that disrupt germline differentiation, such as the loss of Yorkie (Yki), also lead to the loss of EC processes [27]. We used UAS-mCD8RFP in combination with C587-GAL4 to specifically label EC cellular processes with membrane-tethered RFP, mCD8RFP. Similar to the control germarium, EC cellular processes in both mir-124[6] and mir-124[7] mutant germaria spread normally into the germarium to ensheath differentiating germline cells (S7A–S7C’ Fig). This contrasts with the response to expression of UAS-yki RNAi (S7D and S7D’ Fig).

Inhibition of Dpp production from ECs can rescue miR-124 mutant differentiation defects

Genetic alterations of ECs have previously been found to increase BMP signaling by transcriptionally inducing the ectopic expression in ECs of either Dpp or Dally, a glypican that can promote the spread of Dpp signaling [66,67]. Here, the major focus of action of miR-124 is in TF and Cap cells rather than ECs and ectopic BMP signaling is limited to anterior regions. We therefore investigated whether Dpp or Dally production might be altered by miR-124, and in which cells this might occur.

To look for direct evidence of Dpp or Dally induction, we extracted RNA from germaria and early egg chambers of different genotypes for qRT-PCR tests (Fig 3A). We found that dpp RNA levels were 60% higher in miR-124 mutant germaria than in controls, while dally RNA levels were unchanged (Fig 3B). We then used fluorescence-activated cell sorting (FACS) for the expression of UAS-RFP driven by C587-GAL4 to purify ECs prior to qRT-PCR analysis (Fig 3A). We found that dpp RNA levels were increased almost 30-fold for miR-124 mutants compared to controls, while the expression of dally was unchanged (Fig 3C). We conclude that dpp RNA is substantially induced in the ECs of miR-124 mutant germaria.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Ectopic Dpp expression in ECs mediates germline differentiation defects in the miR-124 mutant.

(A) Scheme for collecting control and miR-124-deficient samples by dissociating and sorting RFP-labeled ECs (upper panel) or collecting the germaria of early egg chambers (lower panel). (B) Relative mRNA levels of dpp and dally in the collected germaria of the early egg chambers of miR-124[6] mutants compared with those of the control. Three biological replicates were performed. Data are presented as the mean value ± SEM; significance was determined by two-sided unpaired Student’s t test (* P < 0.05; ns: not significant). (C) Relative mRNA levels of dpp and dally in the sorted ECs of miR-124[6] mutants compared with those of the control. Three biological replicates were performed. Data are presented as the mean value ± SEM; significance was determined by two-sided unpaired Student’s t test (*** P <0.001). (D–F) RNA in situ hybridization of dpp (red) was performed on germaria from wild-type females (D), miR-124[6] (E), and miR-124[7] (F). Dpp-expressing Cap cells are marked by yellow arrowheads, and ECs with dpp signals are marked by white arrowheads. (G) Percentage of the germaria containing the dpp signal in either the Cap cells or ECs of wild-type (w¹¹¹⁸) and 2 independent miR-124 mutant lines, miR-124[6] and miR-124[7]. The number of analyzed germaria is shown above each bar. Significance between each miR-124 mutant line and the control was determined by Fisher’s exact two-sided test (*** P < 0.001). (H–J) C587-GAL4 was used to express UAS-dpp RNAi in ECs of wild-type (H), miR-124[6] (I), or miR-124[7] (J) mutant. Germaria were stained for Hts (red), Vasa (green), and DAPI (blue) to label spectrosomes, fusomes, germline cells, and nuclei, respectively. In (H–J), GSCs are highlighted by dashed circles. Scale bar: 10 μm. (K) Percentage of germaria carrying 5 or more spectrosome-containing cells with different genetic backgrounds. The number of analyzed germaria is shown above each bar. Significance between each transgenic group and the control group (w¹¹¹⁸) was determined by Fisher’s exact two-sided test (*** P < 0.001; ns: not significant). Filled gray circles represent the presence, and empty brown circles represent the absence of a given transgene. The raw data underlying panels B, C, G, and K are available in S1 Data. Dpp, decapentaplegic; EC, escort cell; GSC, germline stem cell.

https://doi.org/10.1371/journal.pbio.3002515.g003

We next examined the mRNA expression pattern of dpp directly by fluorescence in situ hybridization. Consistent with previous reports [28,29], dpp transcripts were detected with high specificity in the Cap cells of all wild-type germaria (Fig 3D). Approximately 15% of germaria (n = 20) also exhibited a dpp signal in 1 or more ECs (Fig 3G), with an average of 0.15 ECs in region 1 (the anterior portion of the germarium) and none in region 2a of control germaria (S4 Table). In contrast, in addition to a signal in Cap cells, 91% of the mir-124[6] mutant germaria (n = 22) and 87% of the mir-124[7] mutant germaria (n = 23) exhibited detectable dpp transcripts in ECs (Figs 3E–3G and S8). Among them, 1.05 ECs (mir-124[6]) and 1.09 ECs (mir-124[7]) on average in region 1 of mutant germaria, and 0.36 (mir-124[6]) and 0.48 (mir-124[7]) on average in region 2a of mutant germaria (S4 Table). Most of the ECs with ectopic dpp expression in region 1 were very close to the anterior of the germarium, in keeping with the location of ectopic germline pMad signals and extra CB-like cells.

We then tested the hypothesis that miR-124 mutants delay germline differentiation by inducing ectopic Dpp expression in ECs using a genetic test. As reported previously [27], dpp and dally knockdown in adult ECs using C587-GAL4 and UAS-dpp RNAi or UAS-dally RNAi had no significant effect on GSC maintenance or differentiation (Figs 3H and S9), allowing us to test their effect in combination with the loss of miR-124 function. We found that limiting dpp expression in ECs (C587 > dpp RNAi) completely suppressed the germline differentiation defect due to miR-124 deletion, whereas the reduction of dally expression had no effect (Figs 3I–3K and S9). Thus, the loss of miR-124 induces a germline differentiation defect by causing ectopic Dpp expression in principally anterior ECs.

miR-124 acts in GSC niche cells to prevent ectopic Dpp production in ECs

Our evidence thus far indicates that miR-124 acts in GSC niche cells to prevent a germline differentiation defect that is caused by ectopic Dpp production in ECs. These deductions rely on the specificity of bab1-GAL4 for adult TF and Cap cells (and not ECs) and C587-GAL4 specificity for ECs (and not TF or Cap cells). These specificities have been thoroughly demonstrated in normal ovaries through the direct examination of GAL4 reporters and a variety of genetic perturbation experiments (S4A–S4B’ Fig) [27,28,68]. However, it is crucial to test also whether these specificities are retained in miR-124 mutant germaria. For example, the loss of miR-124 might cause a full or partial EC-to-Cap cell transformation. To test this specific possibility, we used LamC to identify Cap cells in control, mir-124[6] and mir-124[7] mutant ovaries. In all cases, LamC expression was restricted to endogenous TF, Cap cells, and occasionally some muscle sheath cells but was not present in EC locations (S10A–S10C’ Fig). Additionally, the total number of LamC-positive cells with normal appearance in Cap cell locations was unchanged in mir-124[6] and mir-124[7] mutant germaria (S10D Fig). These results indicate that there were no clear transformations between EC and Cap cells in miR-124 mutant germaria. Equally important, the expression patterns of bab1-GAL4 and C587-GAL4 were unchanged in miR-124 mutant germaria, with no ectopic expression of bab1-GAL4 in ECs or of C587-GAL4 in Cap cells or TF cells (S11 Fig). Hence, we can be confident of the prior results that miR-124 functions in TF or Cap cells, which continue to function as GSC niche cells even in the absence of miR-124, to prevent ectopic Dpp production in ECs.

EGFR is a potential direct target of miR-124 in Cap cells

To understand how miR-124 can act in one cell type to affect another, we sought to identify its direct targets. MiRNAs usually bind to specific sequences within the 3′ UTRs of mRNAs, leading to degradation of the targeted mRNAs and reducing translation into proteins [69]. Several transcriptome studies in both vertebrates and invertebrates have, accordingly, demonstrated the up-regulation of direct targets upon miRNA depletion [70,71]. We therefore used bab1-GAL4>UAS-RFP animals to isolate Cap cells and TF cells from control and mutant ovaries using FACS and compare RNA profiles (Fig 4A). RNA sequencing reads were aligned to the Drosophila genome, and gene expression levels were quantified using the RSEM pipeline. Among the 130 predicted miR-124 targets (TargetScan Fly, http://www.targetscan.org/), we found a group of 57 genes with increased expression in miR-124 mutant Cap and TF cells (Fig 4B and S3 Table). The EGFR gene attracted our attention due to the magnitude of the increase (ranking first). The EGFR signaling pathway has previously been implicated in germline differentiation, but acting positively in ECs [72].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. EGFR is a target of miR-124 in Cap cells.

(A) Scheme for dissociating and sorting RFP-labeled control and miR-124-deficient Cap cells (plus TF cells). (B) Differential expression of the 130 predicted miR-124 targets from the sorted cells of the miR-124[6] mutant vs. control. Among them, 43.8% (57 out of 130, orange part) showed higher expressions. (C) The potential binding site of miR-124 in the 3′ UTR of EGFR is presented. Nucleotide sequences showed the complementarity of miR-124 with the wild-type (WT) or mutant (MUT) site (blue letter) in EGFR on the miR-124 seed-binding region. (D) The pMIR-REPORT-EGFR luciferase construct without (EGFR-3′ UTR^WT) or with (EGFR-3′ UTR^MUT) a mutation site was cotransfected into HEK293 cells with either a negative control miRNA (miRNA-NC) or miR-124, and the corresponding luciferase activity was determined by a dual-luciferase assay. Four biological replicates were performed. Data are presented as the mean ± SEM; significance was analyzed by two-way ANOVA with Sidak’s multiple comparisons test (*** P < 0.001; ns: not significant). (E–H’) Germaria from wild-type females (E–E’), miR-124[6] (F–F’), miR-124[7] (G–G’) mutants, and miR-124[7]; bab1 > UAS-miR-124 (H–H’) were immunostained for an indicator (pERK) of EGFR-MAPK signaling activity. pERK (red) was highly expressed in the ECs (marked by arrowheads) of these different genotypes. In contrast, the pERK in the Cap cells (marked by yellow flat circle and stained for LamC antibody in (E), (F), (G), and (H), green) was undetectable for control females (E–E’) but was strongly expressed in 2 miR-124 mutants (F–G’). Overexpression of miR-124 activity (bab1 > UAS-miR-124) in Cap cells largely suppressed the increased pERK signal in miR-124[7] mutant (H–H’). (E), (F), (G), and (H) show the merging of the 3 channels of pERK, Vasa, and DAPI (blue); (E’), (F’), (G’), and (H’) show pERK stained images in black and white. Scale bar: 10 μm. (I) and (J) Quantitation of EGFR levels in ECs (I) or Cap cells (J) per germarium for the 4 genotypes in (E–H’). Each plot in (I) indicates the mean intensity from 10 ECs for each germarium; each plot in (J) indicates the mean intensity of at least 3 Cap cells within each germarium (n = 20 Drosophila germaria were examined for each group). Data are presented as the mean ± SEM. Significance was determined by Kruskal–Wallis one-way ANOVA with Dunn’s test (*** P < 0.001; ns: not significant). The raw data underlying panels D, I, and J are available in S1 Data. EGFR, epidermal growth factor receptor; TF, terminal filament; UTR, untranslated region.

https://doi.org/10.1371/journal.pbio.3002515.g004

There is a single miR-124 canonical recognition site (CACGGAA) located within the EGFR 3′ UTR (Fig 4C). To test whether this site can mediate the effects of miR-124 on EGFR expression, we performed dual luciferase reporter assays in HEK293 cells (Fig 4C). The 3′ UTR of EGFR (UTR^WT) was cloned into the pMIR-REPORT vector downstream of a firefly luciferase gene, and mutations were introduced into the 3′ UTR (UTR^MUT) to abolish the miRNA target site as a negative control. Compared to the control miRNA sequences (NC, EXIQON 479903–001, USA), the luciferase reporter activity of the EGFR-3′ UTR^WT construct was significantly reduced in the presence of miR-124; moreover, the altered EGFR-3′ UTR^MUT construct abolished the repression (Fig 4D). A similar test was performed in Drosophila S2 cells, directly measuring RNA levels of complete EGFR mRNAs. The qRT-PCR results showed that the transcript levels of EGFR were significantly decreased by miR-124 in EGFR^WT co-transfected cells, but were not changed by miR-124 in EGFR^MUT co-transfected cells (S12 Fig). To investigate whether the target site of miR-124 in EGFR 3′ UTR is conserved, we searched the EGFR 3′ UTR sequences in 20 insect species representing 6 orders, including Diptera (13), Lepidoptera (2), Hymenoptera (2), Coleoptera (1), Hemiptera (1), and Orthopterodea (1). The alignment results showed that the target site of miR-124 in the EGFR 3′ UTR is conserved in many Drosophila species, but is not evident in other insects (S13 Fig). These data confirm that miR-124 can act at a specific site within the 3′ UTR to reduce EGFR expression, and this action is likely conserved among several Drosophila species.

If EGFR is a significant target for miR-124 in TF and Cap cells, we would expect to see increased EGFR expression or activity in miR-124 mutants. Consistent with previous findings [26], we found a strong phosphorylated ERK (pERK) MAP kinase signal, indicative of EGFR signaling pathway activity, in ECs but little or no signal in the TF or Cap cells of normal germaria (Fig 4E and 4E’). However, in miR-124 mutant germaria, there was clearly an increased pERK signal in Cap cells (Fig 4F–4G’). When we expressed UAS-miR-124 in TF and Cap cells in a miR-124 mutant background, the increased pERK signal was rescued to the level in Cap cells of normal germaria (Fig 4H and 4H’). We next measured the average intensity of pERK staining among at least 3 Cap cells and among at least 10 ECs in each germarium. We found a significant elevation of pERK in the Cap cells of miR-124 mutant germaria, while the pERK in ECs appeared unchanged (Fig 4I and 4J). We deduce that miR-124 normally limits EGFR pathway activity in Cap cells and likely does so by directly interacting with the EGFR 3′ UTR to reduce expression.

Altered EGFR signaling as a mediator of miR-124 function in Cap cells

To determine whether EGFR might be a functionally important target for miR-124 in ovaries, we first determined the consequences of overexpressing EGFR in TF and Cap cells by using the combination of UAS-EGFR and bab1-GAL4 compared to controls with each genetic element alone. We found that EGFR overexpression in TF and Cap cells caused a significant germline differentiation defect, with approximately 70% of germaria containing more than 5 cells with a spectrosome (Fig 5A–5C). Similar results were found when we used a temperature-sensitive conditional GAL4/GAL80^ts system to drive UAS-EGFR expression only in adults (S14 Fig). We also found that EGFR overexpression in TF and Cap cells induced ectopic pMad staining, with significantly more pMad-positive cells than in controls (S15 Fig). The locations and magnitudes of the changes in pMad staining and spectrosome-containing cells were very similar for the loss of miR-124 and overexpression of EGFR. EGFR overexpression in TF and Cap cells also significantly increased dpp RNA levels in RNA isolated from the germarium and early egg chambers (Fig 5D). The increase in dpp RNA was 90%, similar to the 60% increase found in analogous miR-124[6] samples that were not enriched for ECs. We next determined the consequences of overexpressing EGFR in ECs by using the combination of UAS-EGFR and C587-GAL4. We found that EGFR overexpression in ECs caused no change in germline differentiation (S16 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. GSC progeny differentiation depends on limiting EGFR signaling activity in Cap cells.

(A and B) Germaria of animals with UAS-EGFR alone (A) or bab1-GAL4 plus UAS- EGFR (B) were stained for Hts (red), Vasa (green), and DAPI (blue), respectively. In (A) and (B), the GSCs are highlighted by dashed circles, whereas the CBs or CB-like in (B) are indicated by arrowheads. Scale bar: 10 μm. (C) Percentage of germaria carrying 5 or more spectrosome-containing cells with different genotypes. The number of analyzed germaria is shown above each bar. Significance was determined by Fisher’s exact two-sided test (*** P < 0.001; ns: not significant). (D) Relative mRNA levels of dpp when EGFR was overexpressed in Cap cells. Three biological replicates were performed. Data are presented as the mean ± SEM; significance was determined by Kruskal–Wallis one-way ANOVA with a Dunn’s test (*** P < 0.001; ns: not significant). In (C) and (D), filled gray circles represent the presence, and empty brown circles represent the absence of a given transgene. (E–G) bab1-GAL4 was used to express UAS-EGFR RNAi in wild-type (E), miR-124[6] (F), or miR-124[7] (G) mutant Cap cells. Germaria were stained for Hts (red), Vasa (green), and DAPI (blue). GSCs are highlighted by a dashed circle in (E), (F), and (G). Reduction of EGFR in Cap cells restored the normal germ cell differentiation for miR-124 mutant lines. Scale bar: 10 μm. (H) Percentage of the germaria carrying 5 or more spectrosome-containing cells with different genotypes. The number of analyzed germaria is shown above each bar. Significance between each transgenic group and the control group (w¹¹¹⁸) was determined by Fisher’s exact two-sided test (*** P < 0.001; ns: not significant). Filled gray circles represent the presence, and empty brown circles represent the absence, of a given transgene. The raw data underlying panels C, D, and H are available in S1 Data. CB, cystoblast; Dpp, decapentaplegic; EGFR, epidermal growth factor receptor; GSC, germline stem cell.

https://doi.org/10.1371/journal.pbio.3002515.g005

Thus, the EGFR gene includes a functional target site for miR-124, EGFR RNA levels were increased in TF and Cap cells by the loss of miR-124, and the overexpression of EGFR in TF and Cap cells, the critical site of action of miR-124, produced phenotypes very similar to those of the loss of miR-124, including dpp RNA induction, ectopic BMP signaling, and defective germline differentiation.

To test whether the miR-124 mutant differentiation defect depends on elevated EGFR activity in TF and Cap cells, we reduced EGFR expression in those cells using bab1-GAL4 and UAS-EGFR RNAi in normal and miR-124 mutant germaria. The reduction of EGFR alone in TF and Cap cells had no effect on germarial appearance or germline differentiation, consistent with the observed very low levels of EGFR signaling pathway activity in these cells. However, the reduction of EGFR expression in TF and Cap cells fully suppressed the differentiation defects of miR-124 mutant animals (Fig 5E–5H). These results indicate that miR-124 normally restricts EGFR expression, likely through direct interaction with the EGFR 3′ UTR, to limit EGFR signaling in TF and Cap cells, thereby preventing ectopic Dpp expression in ECs and delayed germline differentiation.

Notch signaling acts as a connection between the 2 different niches

To find out how the miR-124-EGFR cascade in TF and Cap cells alters dpp transcription in ECs, we searched for potential mediators using the same set of RNASeq results that identified regulation of EGFR by miR-124. We found that the expression of Delta (Dl), the ligand of Notch signaling, was significantly increased in miR-124 mutant cells using the exact test of edgeR based on RNA-seq reads count (TPM = 1.97 in CK, TPM = 20.82 in mutant; P < 0.001) (S17 Fig). Because Dl is a target of EGFR signaling in some settings [73,74], we then asked whether Notch signaling mediates the connection between the miR-124-EGFR cascade in Cap cells and dpp expression in ECs. Consistent with previous findings [75], we found little or no Dl protein expression in Cap cells of normal germaria (Fig 6A and 6A’). However, in miR-124 mutant germaria, there was clearly an increased Dl signal in Cap cells (Fig 6B, 6B’, and 6E). Moreover, reducing EGFR expression in Cap cells significantly suppressed the increased Dl signal of miR-124 mutant animals (Fig 6C, 6C’, and 6E). In addition, overexpression of EGFR alone sufficed to increase Dl expression in Cap cells (Fig 6D, 6D’, and 6E). Collectively, these results suggest that Dl is a downstream target of the miR-124-EGFR cascade in Cap cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Notch signaling acts as a connection between EFGR activity in Cap cells and dpp transcripts in ECs.

(A–D’) Germaria from wild-type females (A–A’), miR-124[6] (B–B’), miR-124[6]; bab1 > UAS-EGFR RNAi (C–C’), and bab1 > UAS-EGFR (D–D’) were immunostained for Dl (the ligand of Notch signaling). Dl (green) showed extremely low expression in Cap cells, marked by yellow flat circle and with mCD8RFP in control females (A–A’) but was more strongly expressed in miR-124 mutants (B–B’) and EGFR-overexpressing germaria (D–D’). In contrast, reduction of EGFR activity (UAS-EGFR RNAi) in Cap cells largely suppressed the increased Dl signal in miR-124 mutants (C–C’). (A), (B), (C), and (D) show the merging of the 3 channels of Dl, RFP, and DAPI (blue); (A’), (B’), (C’), and (D’) show Dl stained images in black and white. Scale bar: 10 μm. (E) Quantitation of Dl protein levels in Cap cells per germarium of flies with different genotypes in (A–D’). Each plot indicates the mean intensity from 3 Cap cells within each germarium (left to right: n = 38, 59, 42, and 57 were examined for each group). Data are presented as the mean ± SEM. Significance was determined by Kruskal–Wallis one-way ANOVA with Dunn’s test (*** P < 0.001; ns: not significant). (F and G) C587-GAL4 was used alone (F) or to express UAS-Notch RNAi (G) in ECs of the miR-124[6] mutant. Germaria of both genotypes were stained for Hts (red), Vasa (green), and DAPI (blue), respectively. In (F) and (G), the GSCs are highlighted by dashed circles, whereas the CBs or CB-like are indicated by arrowheads. Scale bar: 10 μm. (H) Percentage of germaria carrying 5 or more spectrosome-containing cells with different genotypes. The number of analyzed germaria is shown above each bar. Significance was determined by Fisher’s exact two-sided test (*** P < 0.001; ns: not significant). (I) Relative mRNA levels of dpp in the collected germaria of the early egg chambers from different Drosophila in (H). Three biological replicates were performed. Data are presented as the mean value ± SEM; significance was determined by ordinary one-way ANOVA along with Holm–Sidak’s test (*** P < 0.001; ns: not significant). (J) Schematic diagram showing that miR-124 non-cell-autonomously maintains the GSC daughter cell differentiation niche through the EGFR and Notch signaling pathways. In wild-type females (left panel), miR-124 directly down-regulates EGFR in Cap cells, which suppresses Dl expression, thus limiting Notch activity in ECs to prevent BMP ligand production, thereby providing a permissive environment for GSC progeny differentiation. In miR-124 mutants (right panel), higher EGFR activity in Cap cells elevates Dl production, which activates Notch activity to increase Dpp in ECs and cause germline differentiation defects. The raw data underlying panels E, H, and I are available in S1 Data. BMP, bone morphogenetic protein; CB, cystoblast; Dpp, decapentaplegic; EC, escort cell; EGFR, epidermal growth factor receptor; GSC, germline stem cell.

https://doi.org/10.1371/journal.pbio.3002515.g006

To test whether the ectopic dpp expression in ECs of miR-124 mutant depends on elevated Dl protein in Cap cells, we reduced Notch signaling activity in ECs using C587-GAL4 and UAS-Notch RNAi in normal and miR-124 mutant germaria. The reduction of Notch activity alone in ECs had no effect on germline differentiation. However, the reduction of Notch expression in ECs significantly suppressed the differentiation defects of miR-124 mutant animals (Fig 6F–6H). Accordingly, reduction of Notch activity in ECs also significantly decreased dpp RNA levels (Fig 6I). These results support the interpretation that Notch signaling acts directly as a relay between the miR-124-EGFR cascade in Cap cells and dpp transcripts in ECs, leading to a defect in germline differentiation in the absence of miR-124 (Fig 6J). We want to note that the most prevalent sites of ectopic dpp were in anterior ECs (S4 Table), which can contact Cap cells and the inside Notch signaling directly. The less frequent activation of dpp in more posterior ECs (S4 Table) might be due to longer-range Dl-Notch signaling [76,77], potentially mediated by EC processes, or to occasional ectopic Dl production in anterior ECs.

Drosophila stocks

Drosophila stocks were maintained on standard cornmeal/molasses/agar medium within 6-ounce square bottom plastic fly bottles at 25°C. All experiments with wild-type D. melanogaster were carried out with the w¹¹¹⁸ strain. A total of 90 miRNA mutant lines covering 111 Drosophila miRNAs were obtained from Bloomington Drosophila Stock Center (BL) (S1 Table). Other strains were also used in our previous study: C587-GAL4, tub-GAL80^ts, UAS-mCD8-RFP, and UAS-RFP [27]. bab1-GAL4 line (BL#6802) was provided by Prof. Zheng Guo (Huazhong University of Science and Technology, China) [57]. UAS-miR-124 (BL#41126), UAS-dpp RNAi (BL#31530), UAS-dally RNAi (BL#33952), UAS-EGFR (BL#9535), UAS-EGFR RNAi (BL#31183), UAS-Notch RNAi (BL#7078) lines were also obtained from BL. Two independent knockout alleles, miR-124[6] and miR-124[7], and the reporter line miR-124:dsRed were provided by Prof. Eric C. Lai (Sloan-Kettering Institute, New York, New York, United States of America). Characterization of miR-124 loss-of-function alleles in this study is provided in S18 Fig, and their original resources also can be found in Sun and colleagues [49] and Weng and Cohen [50].

Immunohistochemistry

Drosophila ovaries were dissected in 1 × PBS, fixed in 4% paraformaldehyde in PBS for 15 min, rinsed 3 times with 1 × PBST (PBS containing 0.1% Triton X-100 and 0.05% Tween 20), blocked with 1% bovine serum albumin in 1× PBST, and stained with primary antibodies overnight at 4°C. The following antibodies were used: mouse anti-Hts 1B1 [1:20; Developmental Studies Hybridoma Bank (DSHB)], rat anti-Vasa (1:10; DSHB), mouse anti-Lamin C LC28.26 (1:50; DSHB), mouse anti-Delta (1:100; DSHB), rabbit anti-pS423/S425 Smad3 (1:100; Abcam, Cat#52903), rabbit anti-pT202/Y204 ERK1/2 (1:200; Cell Signaling Technology, Cat#4370S). The ovaries were then washed 3 times in 1 × PBST and incubated with Alexa Fluor 488 or 594 secondary antibodies (1:1,000; Molecular Probes) for 2 h at room temperature, followed by 2 washes in 1 × PBST and 1 wash in 1 × PBS and mounting in ProLong Gold Antifade Mountant with DAPI (Invitrogen, Cat#P36935). Fluorescence images were captured using 63 ×/1.4 NA oil-immersion lenses with a Zeiss LSM 800 confocal microscope and then processed using Affinity Designer (Affinity) or ImageJ (National Institutes of Health).

Drosophila fertility assay

Newly emerged male and female w¹¹¹⁸ or miR-124 mutant (miR-124[6] and miR-124[7]) adults were collected. Twenty 2-day-old females and five 2-day-old males of each Drosophila line were placed into a fly bottle, which covered a 3.5-cm diameter Petri dish containing fly food medium. The food dishes were replaced daily for a period of 5 days. The total number of eggs deposited on the dishes each day was counted, and the hatching larvae were counted 2 days after egg collection. Egg hatching rate was calculated as the total number of hatching larvae divided by the number of total eggs laid.

Fluorescent in situ hybridization

Fluorescent RNA in situ hybridization was performed based on the methods as previously described with some modifications [29]. To generate a sequence-specific probe of Drosophila dpp, the target region was amplified from the dpp transcript sequence with an accession number NM_057963.5 using the following primers: 5′-CACCAGAGTTGCAAGCGACCATG-3′ and 5′-CGCACTGTGTTGGCCGACTTG-3′ with a T7 promoter sequence (5′-TAATACGACTCACTATAGGGAGA-3′) at the 5′ extremity. The Dig-labeled probe was synthesized using a T7 High Yield RNA Transcription Kit (Vazyme, Cat#TR101) and digoxigenin nucleotide mixes (DIG-11-UTP, Roche, Cat#11209256910) according to the manufacturer’s instructions. For in situ hybridization, approximately 10 pairs of ovaries were dissected in RNAse-free 1× PBS, separating the ovarioles over the anterior 1/3 of each ovary, and immediately fixed in 4% formaldehyde in 1× PBST at 4°C overnight. The next day, the ovary samples were washed 5 min in 1× PBST 3 times and then treated with proteinase K (50 μg/ml in PBST) for 10 min at room temperature. After removing the proteinase K solution, samples were added to 2 mg/ml glycine solution to divert residual proteinase K activity, washed 3 times with 1× PBST, and refixed in 4% formaldehyde for 30 min, followed by incubation with prehybridization solution (50% formamide, 5× SCC, 0.1% Tween-20, 100 μg/ml salmon sperm DNA, and 100 μg/ml heparin) for 1 h at 56°C. Samples were then hybridized with the dpp probe overnight at 56°C with hybridization buffer. The next day, the ovaries were rinsed for 5 min in 1× PBST at 56°C 4 times, followed by incubation with anti-Dig-POD (1:200; Roche, Cat#11207733910) at room temperature for 30 min. The samples then developed in situ signals using a TSA Fluorescein system (AKOYA Biosciences, Cat#NEL701A001KT). Finally, the ovary samples were mounted in ProLong Gold Antifade Mountant with DAPI (Invitrogen, Cat#P36935), and the fluorescence images were captured using 63 ×/1.4 NA oil-immersion lenses with a Zeiss LSM 800 confocal microscope.

Sample preparation and flow cytometry

C587>RFP and C587>RFP; miR-124[6] mutant flies were used to collect ECs, and bab1>RFP and bab1>RFP; miR-124[6] mutant flies were used to collect TF and Cap cells by using FACS to sort for cells expressing RFP. Specifically, ovaries from 3- to 5-day-old female adults were dissected in sterile 1× PBS. For each group, 100 to 120 pairs of ovaries in total were washed twice in calcium-free 1× PBS and incubated at room temperature for 20 min with intermittent gentle shaking in 700 μl of 0.5% trypsin and 0.25% collagenase in 1× PBS. Then, the dissociated ovaries were filtered through a 40 μm nylon mesh, and the individual cells were collected into 1.5 ml tubes. After centrifugation at 425 × g at 4°C for 5 min, the supernatant was removed, and the cell pellet was resuspended in 150 μl 1× PBS on ice for subsequent cell sorting.

FACS was performed with a MoFlo XDP high-speed cell sorter (ML99030; Beckman Coulter) using Summit 5.0 software (Beckman Coulter). The light source was a 561 nm diode laser (300 mW power), and the cytometer emission filter used was a 580/23 (FL6) to obtain the RFP emission. Cells of each sample were sorted into 10 μl 1× PBS, snap frozen in liquid nitrogen, and then stored at −80°C until use.

Transcriptome analysis

Total RNA of sorted TF and Cap cells was extracted using an RNeasy Mini Kit (QIAGEN, Cat#74106), and purified mRNA was used to synthesize a cDNA library according to the protocol supplied by Illumina. Paired-end reads (150 bp) were obtained from the Illumina HiSeq X Ten platform. Raw sequence data were filtered by removing sequences with low-quality reads, adapter sequences and ambiguous bases and were used for alignment analysis with the D. melanogaster genomic sequences. The transcript expression level was calculated using the RSEM (v1.3.2) pipeline [98]. Differential expression analysis was performed by edgeR v3.30.3 on counts generated by RSEM [99]. Genes with an absolute value of log2 (expression fold change) of ≥1 in the comparison were identified as significantly up-regulated genes in the TF and Cap cells of the miR-124 mutant.

Quantitative real-time PCR

To compare the relative RNA expression levels of dpp and dally in ECs between the w¹¹¹⁸ control and miR-124 mutant flies, we sorted the cells via FACS as described above. Since we could not collect ECs using C587>RFP tools in the germaria with EGFR overexpressed in Cap cells with the help of bab1-GAL4, we alternatively dissected the most anterior part of the Drosophila ovarioles that covered the germarium and the early 3 egg chambers to detect the dpp levels.

Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Cat#74106) and then reverse transcribed into cDNA using HiScript III RT SuperMix for qPCR (Vazyme, Cat#R223-01) according to the manufacturer’s protocol. Quantitative real-time PCR (qRT-PCR) was performed in the QuantStudio3 Real-Time PCR System (Thermo Fisher Scientific) with the ChamQ SYBR qPCR Master Mix Kit (Vazyme, Cat#Q311-02). The primers used in this experiment are listed in S5 Table. Reactions were carried out for 30 s at 95°C, followed by 40 cycles of three-step PCR for 10 s at 95°C, 20 s at 55°C, and 20 s at 72°C. The RNA levels of the target genes were normalized to that of actin 5C mRNA, and their relative concentrations were determined using the 2^-ΔΔCt method [100].

Plasmid construction and luciferase assays

The 130 putative target genes of miR-124 in Drosophila were predicted by TargetScanFly (http://www.targetscan.org/fly/) (S3 Table). In combination with the transcriptome data, the EGFR gene was selected to perform the luciferase assay for target validation due to the highest fold changes (Rank #1) of the up-regulated genes in the TF and Cap cells of the miR-124[6] mutant. The luciferase assay was carried out as described in our previous publication with some modifications [101]. Briefly, the 3′ UTR fragment (863 bp) of EGFR containing the miR-124 target site was PCR-amplified from Drosophila cDNA using specific primers and the 2 × Phanta Max Master Mix Kit (Vazyme, Cat#P525-02). Then, the fragments were cloned into the pMIR-REPORT miRNA Expression Reporter Vector (Ambion, Cat#AM5795) to generate a luciferase-EGFR reporter vector. Mutations at the miR-124 seed sequence binding region were introduced into the miRNA target site in the 3′ UTR of EGFR (GTGCCTT to CACGGAA) using a PCR-based method with specific primers. The mutant construct was used as the negative control. Both the normal (EGFR-3′ UTR^WT) and mutant constructs (EGFR-3′ UTR^MUT) were co-transfected with a synthesized miR-124 (5′-UAAGGCACGCGGUGAAUGCCAAG-3′; Sangon Biotech) or a negative-control miRNA (miRCURY LNA microRNA Mimic Negative Control; Exiqon, Cat#479903–001) in HEK293 cells using Lipofectamine 3000 (Invitrogen, Cat#L3000015), and the resulting luciferase activity of each combination was determined by a dual-luciferase assay at 48 h post-transfection to measure the interaction between miR-124 and EGFR. The Dual-Luciferase Reporter Assay Kit (Vazyme, Cat#DL101-01) was employed according to the manufacturer’s instructions using the SpectraMax iD5 Multi Mode Microplate Reader (Molecular Devices). All primers used for vector construction are listed in S5 Table.

Data analysis and statistics

Statistical analyses were performed in GraphPad Prism version 8.0 (GraphPad Software). Data were analyzed for statistical significance using unpaired two-tailed Student’s t test (Fig 3B and 3C), one-way analysis of variance (ANOVA) along with Dunn’s test (Figs 2H, 4I, 4J, 5D, 6E, S3, S6, S8, S10D, S15D, S15E, and S16B and S2 Table), ordinary one-way ANOVA along with Holm–Sidak’s test (Fig 6I), two-way ANOVA with Sidak’s multiple comparisons test (Figs 4D and S12B), two-way ANOVA with Dunnett’s multiple comparisons test (S2 Fig), and Fisher’s exact two-sided test (Figs 1D, 1G, 2C, 2D, 3G, 3K, 5C, 5H, 6H, S1C, S5C, S9, S14, and S16A). Error bars indicate the standard error of the mean (SEM), and the datasets are represented as the mean ± SEM. Significance values are indicated as follows: * P < 0.05, ** P < 0.01, and *** P < 0.001.