1. Stem Cells and Regenerative Medicine
Download icon

Wnt signaling-mediated redox regulation maintains the germ line stem cell differentiation niche

  1. Su Wang
  2. Yuan Gao
  3. Xiaoqing Song
  4. Xing Ma
  5. Xiujuan Zhu
  6. Ying Mao
  7. Zhihao Yang
  8. Jianquan Ni
  9. Hua Li
  10. Kathryn E Malanowski
  11. Perera Anoja
  12. Jungeun Park
  13. Jeff Haug
  14. Ting Xie  Is a corresponding author
  1. Stowers Institute for Medical Research, United States
  2. University of Kansas School of Medicine, United States
  3. Tsinghua University, China
Research Article
  • Cited 19
  • Views 1,565
  • Annotations
Cite this article as: eLife 2015;4:e08174 doi: 10.7554/eLife.08174

Abstract

Adult stem cells continuously undergo self-renewal and generate differentiated cells. In the Drosophila ovary, two separate niches control germ line stem cell (GSC) self-renewal and differentiation processes. Compared to the self-renewing niche, relatively little is known about the maintenance and function of the differentiation niche. In this study, we show that the cellular redox state regulated by Wnt signaling is critical for the maintenance and function of the differentiation niche to promote GSC progeny differentiation. Defective Wnt signaling causes the loss of the differentiation niche and the upregulated BMP signaling in differentiated GSC progeny, thereby disrupting germ cell differentiation. Mechanistically, Wnt signaling controls the expression of multiple glutathione-S-transferase family genes and the cellular redox state. Finally, Wnt2 and Wnt4 function redundantly to maintain active Wnt signaling in the differentiation niche. Therefore, this study has revealed a novel strategy for Wnt signaling in regulating the cellular redox state and maintaining the differentiation niche.

https://doi.org/10.7554/eLife.08174.001

eLife digest

An animal or plant has many different types of cells that have specific roles in the life of the organism. These cells are organized into tissues. In most tissues in adult animals, small groups of cells called stem cells are responsible for replacing the other cells that have been lost due to disease, injury, or as part of normal body maintenance.

The ‘germ line’ stem cells of female fruit flies—which produce female sex cells (or eggs)—are an effective system for studying how stem cells are regulated. These cells live in an area of the ovary called a stem cell niche. Each time a stem cell divides, it produces one stem cell and one other daughter cell. This daughter cell then moves into another niche called the ‘differentiation’ niche and undergoes a series of divisions that produce the egg cells. The differentiation niche is formed by escort cells and is crucial for producing the egg cells, but it is not clear how the escort cells promote this process, or how the niche is maintained.

Wang et al. have now studied the differentiation niche in more detail. The experiments show that a cell communication system called Wnt signaling maintains the differentiation niche by controlling the ability of the escort cells to grow and divide. If Wnt signaling is defective, the differentiation niche is lost, which disrupts the formation of egg cells.

Further experiments show that two proteins called Wnt2 and Wnt4 in the differentiation niche—which activate Wnt signaling—act as signals to regulate the niche, mainly by controlling the expression of four particular genes. These four genes encode enzymes that remove ‘reactive oxygen species’ from cells. Wang et al.'s findings have revealed an important role for Wnt signaling in maintaining the differentiation niche. The next step is to figure out the details of how this works.

https://doi.org/10.7554/eLife.08174.002

Introduction

Stem cells have two important properties, self-renewal and differentiation, which are critical for continuously generating new functional cells to maintain tissue homeostasis. The self-renewal property is controlled in various stem cell systems by interplays between signals from the niche and intrinsic factors (Li and Xie, 2005; Morrison and Spradling, 2008; Losick et al., 2011). Germ line stem cells (GSCs) in the Drosophila ovary and testis are attractive systems for studying stem cell self-renewal at the molecular and cellular level (Fuller and Spradling, 2007; Xie, 2013). Although stem cell differentiation was widely thought to be a developmentally default state, we have recently proposed that GSC lineage differentiation is also controlled extrinsically by a differentiation niche formed by inner germarial sheath cells (ISCs, also known as escort cells). However, it remains unclear how the maintenance and function of the differentiation niche are regulated at the molecular level. In this study, we show that autocrine Wnt2/4 signaling maintains the differentiation niche by regulating ISC proliferation and survival via redox regulation.

In the Drosophila ovary, two or three GSCs at the tip of the germarium, the most anterior region of the Drosophila ovary, continuously self-renew and generate differentiated GSC daughters, cystoblasts (CBs). The CBs further divide four times synchronously with incomplete cytokinesis to form 2-cell, 4-cell, 8-cell, or 16-cell cysts (de Cuevas et al., 1997). GSCs and their differentiated progeny can be reliably identified by their unique morphology of germ line-specific intracellular organelles known as fusomes: GSCs and CBs contain a spherical fusome known as the spectrosome, whereas differentiated germ cell cysts contain a branched fusome (Lin et al., 1994). GSCs can be reliably distinguished from CBs by their direct contact with cap cells (Figure 1A). Cap cells function as the self-renewing niche to maintain GSCs by activating BMP signaling and maintaining E-cadherin-mediated cell adhesion (Song et al., 2002; Xie and Spradling, 1998, 2000). In addition, various classes of intrinsic factors work with BMP signaling and E-cadherin to control GSC self-renewal (Xie, 2013). Therefore, GSC self-renewal is controlled by coordinated functions of niche-initiated signaling pathways and intrinsic factors.

Figure 1 with 1 supplement see all
Canonical Wnt signaling in ISCs promotes germ cell differentiation.

(A) The Drosophila germarium dividing into three regions 1, 2a, 2b and 3. Abbreviations: TF-terminal filament; CPC-cap cell; ISC-inner germarial sheath cell; FC-follicle cell; GSC-germ line stem cell; CB-cystoblast; DC-developing cyst; SS-spectrosome; FS-fusome. In BL, cap cells are highlighted by broken ovals, whereas CBs and cysts are indicated by arrowheads and arrows, respectively. (B) In the c587>>UAS-GFP germarium containing two GSCs (spectrosomes indicated by arrowheads) close to cap cells, one CB and a few differentiated cysts are surrounded by GFP-positive ISCs. (CE) In armKD1 (C) dshKD1 (D) germaria, many spectrosome-containing CBs accumulate far away from cap cells. (E) Quantification results on the percentages of the germaria exhibiting the germ cell differentiation defect (≥4 CBs). (FH) fz fz2 double knockdown (F), axn- (G), and sgg-overexpressing (H) germaria contain excess CBs. (IL) In arm*-overexpressing control (I) and dshKD (K, L) germaria, GSC progeny differentiate into cysts containing a branched fusome (arrow). J: Quantification results.

https://doi.org/10.7554/eLife.08174.003

Following GSC division, differentiating GSC daughters, CBs, are always positioned away from the self-renewal niche. ISCs sit on the surface of the germarium to send their cellular processes to wrap up underneath CBs, mitotic cysts, and early 16-cell cysts, which move posteriorly (Decotto and Spradling, 2005; Kirilly et al., 2011; Morris and Spradling, 2011). Our recent study suggests ISCs and their associate long cellular processes act as the differentiation niche to promote GSC progeny differentiation in the Drosophila ovary because disrupting long ISC processes leads to an accumulation of CB-like cells, indicative of a germ cell differentiation defect (Kirilly et al., 2011). A series of genetic studies have further supported the existence of the differentiation niche.

The epidermal growth factor (EGF) signaling pathway is active in ISCs to promote GSC lineage differentiation partly by repressing dally expression (Schultz et al., 2002; Liu et al., 2010). In addition, Rho signaling is also required in ISCs to promote GSC differentiation partly by repressing dally and dpp expression. dally encodes a proteoglycan protein, which is capable of promoting Dpp/BMP diffusion to the differentiation niche (Guo and Wang, 2009; Hayashi et al., 2009). Ecdysteroid signaling also operates in ISCs to promote germ cell differentiation because inactivating ecdysteroid receptors EcR and Usp in ISCs disrupts cyst formation (Morris and Spradling, 2012). One potential mechanism is that ecdysteroid signaling controls the formation of ISC cellular processes, thereby promoting the interaction between ISCs and germ cells (Konig and Shcherbata, 2015). Gap junction protein Inx2 functions in ISCs to promote germ cell differentiation, but its transmitted substances between ISCs and germ cells remain identified (Mukai et al., 2011). The importance of gap junctions between ISCs and germ cells could also explain why ISC cellular processes are important for germ cell differentiation. Therefore, physical interactions and signaling-mediated communications between ISC cellular processes and GSC progeny likely contribute to GSC progeny differentiation collectively.

In addition, chromatin regulators are also important in ISCs to promote GSC differentiation. Eggless, a Drosophila H3K9 trimethyltransferase, maintains ISCs and represses dally and dpp expression in ISCs, thereby promoting germ cell differentiation (Wang et al., 2011). Similarly, Piwi functions in ISCs likely as a chromatin regulator to control germ cell differentiation partly by repressing dpp expression (Jin et al., 2013; Ma et al., 2014). dBre1 (a E3 ubiquitin ligase) and dSet1 (a H3K4 trimethylase) together control H3K4 trimethylation in ISCs and promote germ cell differentiation partly by limiting BMP signaling from the differentiation niche (Xuan et al., 2013). The potential chromatin factor without children (Woc) maintains the ISC-germ cell physical interaction via regulation of Stat-Zfh1 (Maimon et al., 2014). The histone demethylase Lsd1 is required in ISCs to promote germ cell differentiation by maintaining ISC survival, maintaining ISC morphology, and preventing BMP signaling from the differentiation niche (Eliazer et al., 2011, 2014). Therefore, ISCs function as the differentiation niche by preventing BMP signaling and direct communication.

A recent study showed that tyrosine kinase Btk29A maintains Wnt signaling in ISCs by phosphorylating Drosophila β-catenin homolog Armadillo (Arm) (Hamada-Kawaguchi et al., 2014). It also argues that Wnt4 activates Wnt signaling to maintain Piwi expression and repress E-cadherin expression in ISCs, thereby promoting germ cell differentiation. In contrast, this study has demonstrated that both ISC-expressing Wnt2 and Wnt4, but not Wnt4 alone, act through known Wnt pathway components to maintain active Wnt signaling, promoting germ cell differentiation. More importantly, we show that Wnt signaling is required to maintain ISCs by promoting ISC survival and proliferation. Surprisingly, Wnt signaling is dispensable for Piwi expression. Instead, we demonstrate that Wnt signaling controls the expression of four Gst genes to maintain the reduced redox, thereby promoting ISC maintenance and germ cell differentiation. Finally, knocking down one of the Gst gene, GstD2, in ISCs leads to the germ cell differentiation defect. Therefore, our study has revealed a novel mechanism which autocrine Wnt signaling utilizes to maintain ISCs and promote germ cell differentiation.

Results

Canonical Wnt signaling is required in the differentiation niche to promote germ cell differentiation

To identify the genes that function in ISCs to promote germ cell differentiation, we used c587-driven UAS-RNAi expression to knockdown individual genes in ISCs. c587 is a GAL4 line that is specifically expressed in ISCs and early follicle cell progenitors based on UAS-GFP expression (Song et al., 2004) (Figure 1B). To facilitate the identification of GSCs and differentiated germ cells, spectrosomes and fusomes are labeled by Hts staining (Lin et al., 1994), and germ cells are visualized by Vasa staining (Lasko and Ashburner, 1988). In contrast to the control germarium containing 0 to 2 CBs, knocking down Wnt downstream genes armadillo (arm) and disheveled (dsh) in ISCs leads to an accumulation of many spectrosome-containing CBs (collectively referred to single germ cells at least one cell diameter away from cap cells) (Figure 1C,D). Based on the fact that control germaria rarely contain three CBs, the germaria containing four or more CBs are considered to exhibit a germ cell differentiation defect. Over 90% of arm and dsh knockdown germaria (armKD1 and dshKD1) exhibit the germ cell differentiation defect (Figure 1E). Similarly, c587-driven expression of the RNAi lines against different arm or dsh sequences (armKD2 and dshKD2) also generates the similar germ cell differentiation defect (Figure 1E). In Drosophila, Wnt ligands bind the receptor complex composed of Frizzled (Fz), Frizzled 2 (Fz2) and Arrow to activate Dsh and stabilize Arm, which forms a protein complex with a TCF (T Cell Factor)-like Pangolin in the nucleus to activate target gene expression, whereas Axin (Axn) and Shaggy (Sgg) negatively modulate Wnt signaling by promoting Arm degradation (Logan and Nusse, 2004). Simultaneous knockdown down of both fz and fz2 (fzKD+fz2KD) in ISCs recapitulate the germ cell differentiation defect caused by either armKD or dshKD, although either fzKD or fz2KD yields a much weaker germ cell differentiation defect (Figure 1E,F; Figure 1—figure supplement 1). Consistently, overexpression of axn and sgg in ISCs also leads to the germ cell differentiation defect similar to that caused by armKD and dshKD (Figure 1E,G,H). These results indicate that Wnt signaling is required in ISCs to promote germ cell differentiation.

The expression of a constitutively active mutant armS10 (arm*) leads to hyperactive Wnt signaling independently of Wnt ligands (Pai et al., 1997). c587-driven arm* expression alone does not cause any obvious GSC maintenance and germ cell differentiation defects, but results in severe egg chamber budding defects likely due to defective follicle cell development (Figure 1I). Interestingly, c587-mediated arm* expression can fully rescue the germ cell differentiation defect caused by the two independent dshRNAi knockdowns, indicating that Arm functions downstream of Dsh in ISCs to promote germ cell differentiation (Figure 1J–L). These results further suggest that the two dshRNAi lines are highly specific. Taken together, our findings demonstrate that canonical Wnt signaling works in ISCs to promote GSC lineage differentiation.

Wnt signaling is required in adult ISCs to promote GSC lineage differentiation

The c587-gal4 driver is known to be expressed in somatic precursor cells in the developing female gonad, which give rise to terminal filament, cap cells, adult ISCs, and follicular stem cells (Zhu and Xie, 2003). To determine if adult ISCs require active Wnt signaling for GSC lineage differentiation, we took advantage of c587-deriven expression of a temperature-sensitive gal4 repressor gal80 (c587>>UAS-gal80ts) to allow the expression of UAS-RNAi lines against dsh and arm only in adult ISCs using temperature shift. At 25°C, gal80ts is functional to prevent gal4-driven expression of a UAS transgene, but at 29°C, gal80ts is inactivated to allow gal4-driven gene expression (McGuire et al., 2003). After the c587>>gal80ts control and c587>>gal80ts UAS-dshRNAi or UAS-armRNAi females eclosed at 25°C (RNAi expression is extremely low or not expressed), they either continued to be kept at 25°C for one week (keeping RNAi expression repressed) or were moved to 29°C for one week (turning on the RNAi expression due to gal80 inactivation). For the c587>>gal80ts control females, which were cultured at 25°C or 29°C as adults for one week, their germaria contain the normal number of CBs (Figure 2A,B). Interestingly, for the c587>>gal80ts UAS-dshRNAi or UAS-armRNAi females, which were cultured at constant 25°C, their germaria contain the normal or close to normal numbers of CBs (Figure 2B–F). In contrast, for the c587>>gal80ts UAS-dshRNAi or UAS-armRNAi females, which were shifted from 25°C to 29°C for one week, their germaria contain excess CBs, indicative of the germ cell differentiation defect (Figure 2B,C′–F′). These results demonstrate that Wnt signaling is required in adult ISCs to promote germ cell differentiation.

Wnt signaling is required in adult ISCs to promote germ cell differentiation.

Broken ovals highlight germ line stem cells (GSCs), whereas arrowheads indicate cystoblasts (CBs). Two experimental regimens 25–25 and 25–29 mean the females cultured under 25°C or 29°C for one additional week after reaching adulthood at 25°C, respectively. (A, A′) Control c587>>gal80ts germaria contain one or two CBs under the 25–25 (A) or 25–29 (A′) condition. (B) Quantification results on the percentages of the germaria carrying 4 or more CBs show that adult stage-specific arm or dsh knockdown causes an accumulation of more CBs in comparison with the control. (CF) armRANi- (C, D) or dshRNAi- (E, F) carrying germaria contain one or two CBs under 25–25, indicating that RNAi-mediated arm or dsh knockdown in adult inner germarial sheath cells (ISCs) under 25°C is very limited based on the germ cell differentiation phenotype. (C′F′) armRANi- (C′, D′) or dshRNAi- (E′, F′) carrying germaria contain many more CBs under 25–29, indicating that RNAi-mediated arm or dsh knockdown under 29°C in adult ISCs leads to the severe germ cell differentiation defects.

https://doi.org/10.7554/eLife.08174.005

Wnt signaling maintains ISCs by promoting proliferation and survival

Our previous studies have shown that a severe ISC loss also causes the similar germ cell differentiation defect (Kirilly et al., 2011; Wang et al., 2011). Then, we determined if Wnt signaling controls ISC maintenance by using the PZ1444 reporter to quantify ISC numbers in control and Wnt signaling-defective germaria. PZ1444 expresses nuclear LacZ in ISCs and cap cells, which can be distinguished based on nucleus size and location (Xie and Spradling, 2000). In the one-week-old control germaria, PZ1444 labels 30–35 ISCs in addition to cap cells (Figure 3A,C). In the one-week-old arm and dsh knockdown germaria, only fewer than 5 ISCs remain (Figure 3B–D). Some knockdown germaria have completely lost ISCs (Figure 3B), whereas others retain one or a few ISCs (Figure 3D). Consistently, overexpression of axn and sgg also leads to a severe ISC loss (Figure 3E,F). These results demonstrate that Wnt signaling is required to maintain ISCs.

Figure 3 with 1 supplement see all
Wnt signaling maintains the differentiation niche by promoting cell proliferation and survival.

Cap cells are highlighted by broken ovals. (A) The control germarium contains PZ1444-labeled cap cells and ISCs (one by arrow). (BD) The armKD (B) and dshKD (D) germaria maintain zero and one ISC (arrow), respectively. (C) Quantification results on ISC numbers (mean ± standard deviation; Student's t-test is used to calculate p values). (E, F) c587-directed axn- (E) and sgg- (F) overexpressing germaria contain no ISCs (E) and a few ISCs (F, arrows), respectively. (G) c587-directed arm* expression drastically increases the ISC number (one by arrow). (HL) c587-mediated knockdown of arm (I) and dsh significantly decreases BrdU-positive PZ1444-labeled ISCs (arrows), whereas c587-directed arm* expression (K) significantly increases BrdU-labeled ISCs in comparison with the control (H). The arrowhead in I indicates a BrdU-positive germ cell cyst. J and L show quantification results. (MQ) c587-mediated knockdown of arm and dsh (N) significantly increases TUNEL-positive PZ1444-labeled ISCs (arrows), whereas c587-directed arm* expression (P) significantly decreases TUNEL-positive ISCs in comparison with the control (M). (O, Q) Quantification results.

https://doi.org/10.7554/eLife.08174.006

Wnt signaling maintains ISCs possibly by promoting cell proliferation, survival, or both. Interestingly, arm*-expressing germaria contain significantly more ISCs (Figure 3G). In contrast with the one-week-old control germaria containing 32 ISCs, the one-week-old arm*-expressing germaria carry 130 ISCs (Figure 3C). Consistently, c587-mediated axn and sgg knockdown, which increases Wnt-signaling activity (Heslip et al., 1997; Willert et al., 1999), also leads to more ISCs (Figure 3—figure supplement 1). Interestingly, germ cell differentiation proceeds normally in the germaria carrying extra ISCs, suggesting that ISCs provide a permissive environment for germ cell differentiation (Figure 3G; Figure 3—figure supplement 1).

To further investigate how Wnt signaling maintains the ISC population, we examined ISC proliferation and apoptosis in the control, armKD, dshKD, and arm*-expressing germaria using BrdU incorporation and TUNEL (Terminal deoxynucleotidyl transferase dUTP nick end labeling)-labeling assays, respectively. In order to avoid severe ISC loss in these experiments, we purposely cultured the control, armKD, dshKD, and arm*-expressing females at 29°C for shorter time than earlier experiments. BrdU incorporation labels replicating ISCs in the S-phase of the cell cycle, whereas TUNEL labeling detects fragmented DNA in dying cells. Based on BrdU-labeling results, defective Wnt signaling significantly decreases ISC proliferation, whereas hyperactive Wnt signaling significantly increases ISC proliferation (Figure 3H–L). Based on TUNEL-labeling results, defective Wnt signaling significantly increases ISC apoptosis, whereas hyperactive Wnt signaling significantly decreases ISC apoptosis (Figure 3M–Q). These results suggest that Wnt signaling maintains the ISC population by promoting proliferation and increasing survival.

Wnt signaling is required in ISCs to prevent BMP signaling in differentiated GSC progeny and maintain long ISC cellular processes

Our previous studies have also shown that ISC loss disrupts germ cell differentiation by increasing BMP signaling (Kirilly et al., 2011; Wang et al., 2011; Ma et al., 2014). In the Drosophila germarium, BMP signaling leads to production of phosphorylated Mad (pMad) and activation of Dad-lacZ reporter expression in GSCs (Chen and McKearin, 2003; Kai and Spradling, 2003; Casanueva and Ferguson, 2004; Song et al., 2004). In the control germaria, pMad and Dad-lacZ expression is restricted to GSCs (Figure 4A,A′,D,D′). In the armKD and dshKD germaria, pMad and Dad-lacZ expression is activated in the accumulated CBs locating one to a few cells away from cap cells in addition to GSCs (Figure 4B,B′,C,C′,E,E′,F,F′). These results indicate that BMP signaling is spread to the differentiation zone in the Wnt signaling-defective germarium.

Wnt signaling in ISCs prevents BMP signaling in the differentiated germ cell zone and maintains long ISC cellular processes.

Ovals indicate GSCs, whereas arrows in A-F′ denote CBs or CB-like single germ cells. (A′F′) Images only show green fluorescence of (AF). (A, A′) In the control germarium, GSCs are positive for pMad, but one CB is negative. (BC′) In the armKD (B, B′) and dshKD (C, C′) germaria, GSCs are pMad-positive as in the control. Although most of the accumulated CBs are negative for pMad, some CBs are pMad-positive (arrows). (D, D′) The control germarium shows that GSCs are positive for Dad-lacZ expression, but one CB is negative. (EF′) In the armKD (E, E′) and dshKD (F, F′) germaria, GSCs are Dad-lacZ-positive as in the control. Some of the accumulated CBs are negative for Dad-lacZ, but the other ones exhibit low Dad-lacZ expression (arrows). (G) RNA sequencing (RNA-seq) results on the purified ISCs show that mRNA levels for the known BMP pathway components, including dpp, gbb (also encoding a BMP ligand) and dally, remain largely unchanged in axn-overexpressing ISCs or dskKD ISCs in comparison with the control (FKPM = reads per kilobase per million mapped reads). (H, H′) Our CB and GSC quantification results indicate that the heterozygous dpp mutations have partial suppression on the germ cell differentiation defects caused by dshKD, and the suppression is not due to the changes in GSC numbers (50 germaria examined for each genotype). (IK′) armKD (J, J′) and dshKD (K, K′) ISCs lack their CD8GFP-positive cellular processes extending into the accumulated CBs (arrowheads) in contrast with the control ISCs extending the cellular processes wrapping up underneath germ cells (arrowheads, I′). (I′K′) A higher magnification of highlighted areas in (IK).

https://doi.org/10.7554/eLife.08174.008

To investigate how Wnt signaling regulates germ cell differentiation at the molecular level, we compared the gene expression changes in fluorescence-activated cell sorting (FACS)-purified GFP-labeled control, Wnt signaling-defective dsh knockdown, and axn-overexpressing (AxnOE) ISCs using deep RNA sequencing (RNA-seq). Our RNA-seq results show that known Drosophila BMP-signaling components and regulators, including dally and dpp, which are often upregulated in defective ISCs (Liu et al., 2010; Kirilly et al., 2011; Wang et al., 2011; Jin et al., 2013; Ma et al., 2014), remain unchanged in the dshKD and AxnOE ISCs in comparison with the control ISCs (Figure 4G). To further determine if upregulated BMP signaling is responsible for the germ cell differentiation defect resulted from defective Wnt signaling, we determined the effect of heterozygous dpp mutations on the differentiation defect caused by dshKD because they have been shown to suppress the differentiation defect caused by BMP-signaling upregulation (Kirilly et al., 2011; Wang et al., 2011). Our results indicate that the dpp heterozygous mutation dpphr4 significantly rescues the germ cell differentiation defect caused by dshKD1 and dshKD2, whereas the dpp heterozygous mutation dpphr56 significantly rescues that caused by dshKD1 but not dshKD2 (Figure 4H). This rescue effect by the dpp heterozygous mutations is unlikely caused by decreasing GSCs (Figure 4H′). These results suggest that Wnt signaling in ISCs regulates germ cell differentiation partly by preventing BMP signaling in the differentiation niche.

Long ISC cellular processes are also required to promote germ cell differentiation (Kirilly et al., 2011). They can be easily visualized by c587-driven expression of membrane-tethered GFP, CD8GFP. In the control germaria, ISC cellular processes wrap up differentiated germ cells (Figure 4I,I′). In contrast, cellular processes in the remaining ISCs of the armKD and dshKD germaria fail to extend into the accumulated CBs (Figure 4J–K′). These results indicate that Wnt signaling is required to maintain ISC cellular processes.

Wnt signaling is dispensable for maintaining Piwi protein expression in ISCs

A recent study proposes that Wnt signaling upregulates piwi expression in ISCs, thereby promoting germ cell differentiation (Hamada-Kawaguchi et al., 2014). To verify if Piwi protein is indeed downregulated in Wnt signaling-defective ISCs, we quantified the expression of Piwi protein in the PZ1444-labeled dshKD or armKD ISCs. Piwi protein shows comparable expression levels among the examined 203 control ISCs, 48 armKD1 ISCs, 58 armKD2 ISCs, and 46 dshKD2 ISCs (Figure 5A–D). Surprisingly, Piwi protein levels are significantly elevated in the dshKD1 ISCs (93 examined) in comparison with the control ISCs (Figure 5D). In addition, we also used FACS to purify the GFP-labeled control ISCs, axnOE ISCs, and dshKD1 ISCs for RNA-seq. The RNA-seq results show that the piwi mRNA levels are slightly upregulated in the axnOE ISCs and are significantly upregulated in the dshKD ISCs in comparison to the control ISCs (Figure 5E). Therefore, our results indicate that Wnt signaling promotes germ cell differentiation unlikely by maintaining Piwi expression in ISCs. However, Wnt signaling possibly works with Piwi in an unknown means to promote germ cell differentiation because the loss of their functions in ISCs leads to similar germ cell differentiation defects (Jin et al., 2013; Hamada-Kawaguchi et al., 2014; Ma et al., 2014).

Piwi mRNA and protein expression remain normally expressed in the Wnt signaling-defective ISCs.

PZ1444 is used to highlight ISCs. (A) The control germarium shows that ISCs (arrow) express more Piwi proteins than cap cells (oval) and underneath germ cells. (B, C) The remaining armKD (B) and dshKD (C) ISCs (arrows) still retain high Piwi protein expression. (D) Quantification results show that Piwi protein levels remain largely unchanged in the remaining armKD1, armKD2, and dshKD2 ISCs in comparison with the wild-type control ISCs, but it appears to increase its expression in dshKD1 ISCs. (E) RNA-seq results on the purified ISCs show that piwi mRNA expression levels are not changed in axn-overexpressing ISCs, but increase in dskKD ISCs, in comparison with the wild-type control ISCs.

https://doi.org/10.7554/eLife.08174.009

Wnt signaling maintains the cellular redox state in ISCs, GSCs, and early differentiated germ cells by maintaining glutathione-S-transferase (Gst) gene expression

Our RNA-seq results also show that four Gst genes, GstD2, GstD4, GstD10, and GstE3, decrease their mRNA expression levels significantly in both dshKD- and axn-overexpressing ISCs in comparison with the control (Figure 6A). GST proteins comprise a family of eukaryotic metabolic enzymes for eliminating hydrogen peroxide (H2O2) and catalyzing the conjugation of the reduced form of glutathione (GSH) to oxidized substrates for the purpose of detoxification. Thus, downregulated expression of the Gst genes in the Wnt signaling-defective ISCs might cause the increase in cellular reactive oxygen species (ROS), which can be detected by dihydroethidium (DHE) fluorescence. In the anterior half of control germaria, ISCs and underneath differentiating germ cells show low DHE fluorescence, indicating that these cells have low cellular ROS levels, including ISCs (Figure 6B; Figure 6—figure supplement 1A,B′). In contrast, dshKD and armKD ISCs elevate DHE fluorescence, and germ cells underneath also increase DHE fluorescence, further supporting the idea that defective Wnt signaling results in the increased cellular ROS in ISCs and their interacting germ cells (Figure 6B′; Figure 6—figure supplement 1C–G). As a distinct GST family member from the GST-D family members, GST2 has similar function in regulating the cellular redox state (Singh et al., 2001). Consistently, GST2 overexpression in the dshKD ISCs can also restore low DHE fluorescence in the anterior half of the germaria, suggesting that Wnt signaling controls the cellular redox state in ISCs, GSCs, and early GSC progeny by regulating Gst gene expression (Figure 6B″; Figure 6—figure supplement 1H). Catalase (CAT) can also help eliminate cellular H2O2 by converting it to free oxygen and water. As expected, CAT overexpression in the dshKD ISCs also restores low cellular ROS in the anterior half of the germaria (Figure 6B′″; Figure 6—figure supplement 1I). These results suggest that Wnt signaling regulates the cellular redox state in ISCs by controlling Gst gene expression.

Figure 6 with 2 supplements see all
Wnt signaling maintains the reduced redox state in ISCs, promoting germ cell differentiation.

(A) RNA-seq results of the purified ISCs show that GstD2, GstD4, GstD10, and GstE3 mRNA expression levels are significantly lower in axn-overexpressing or dskKD ISCs than the control ISCs. In BB′″, DI, LN, and PR, cap cells and GSCs are highlighted by broken ovals, some CBs are indicated by arrowheads and some PZ1444-positive ISCs are denoted by arrows. (BB′″) The dshKD germarium (B′) exhibits a drastic increase of dihydroethidium (DHE) fluorescence in the anterior region, including ISCs, GSCs, and early GSC progeny, in comparison with the control germarium (B). GST2 (B″) or catalase (CAT) (B′″) overexpression in ISCs restores low DHE fluorescence in the dshKD germariun. (CJ) GST2 (D, G), CAT (E, H), or superoxide dismutase1 (SOD1) (F, I) overexpression in dshKD ISCs significantly decreases CBs (arrowheads; DF) and significantly increase ISCs (arrows; GI) in comparison with dshKD. (C, J) Quantification results on CB and ISC numbers, respectively (for each genotype, 50 or more germaria examined). (K) Quantitative RT-PCR results on the purified ISCs show that GstD2 knockdown by RNAi Line 1 (GstD2KD1), but not the line 2 (GstD2KD2), significantly decreases GstD2 mRNA levels in comparison to the control (c587). (LO) GstD2KD1 (L) and CatKD germaria contain 4 CBs and 0 CB, respectively, but the GstD2KD1 CatKD germarium (N) contains 5 CBs (O: CB quantitative results). (PS) GstD2KD1 CatKD germarium (R) contains fewer ISCs (arrows) than GstD2KD1 (P) and CatKD (Q) germaria (S: ISC quantitative results). Note: PZ1444 expression appears to be downregulated in GstD2KD1 CatKD ISCs.

https://doi.org/10.7554/eLife.08174.010

Then, we determined if the restoration of the reduced redox state in dshKD ISCs could rescue the germ cell differentiation defect. Cytoplasmic superoxide dismutase1 (SOD1) is also important for the clearance of cellular ROS. The germaria overexpressing GST2, CAT, and SOD1 in ISCs contain normal CB numbers (Figure 6C; Figure 6—figure supplement 2A–C). GST2, CAT, and SOD1 overexpression in the dshKD ISCs drastically and significantly reduces CB numbers in the germaria, but does not change GSC numbers significantly, indicating that increased cellular ROS levels in dshKD ISCs are largely responsible for the germ cell differentiation defect (Figure 6C–F; Figure 6—figure supplement 2D–G). Similarly, GST2, CAT, and SOD1 overexpression in the dshKD ISCs significantly and drastically rescue the ISC number, but not to the wild-type ISC number, indicating that increased ROS is at least partly responsible for the loss of dshKD ISCs (Figure 6G–J; Figure 6—figure supplement 2H–M). Taken together, these results demonstrate that Wnt signaling in ISCs maintains the reduced redox state, which is partly responsible for ISC maintenance and germ cell differentiation.

GstD2 works with Cat in ISCs to control germ cell differentiation

Our RNA-seq results indicate that GstD2 is the most abundantly expressed Gst genes in ISCs. In addition, it is also the most severely downregulated Gst gene in the Wnt signaling-defective ISCs, which prompted us to further investigate its function in promoting germ cell differentiation (Figure 6A). We generated two independent RNAi lines against GstD2, among which Line 1 (GstD2KD1) efficiently knocks down GstD2 mRNA but Line 2 (GstD2KD2) does not based on quantitative RT-PCR results on the purified ISCs (Figure 6K). In contrast with the control germaria containing one CB on average, the GstD2KD1 germaria contain two CBs, which are significantly more than the control (Figure 6L,O). As expected, the GstD2KD2 germaria behave like the control, containing one CB on average (Figure 6O). Interestingly, c587-mediated expression of Gst2 and Cat can also rescue the moderate germ cell differentiation defect caused by GstD2KD1, indicating that the differentiation defect is caused by GstD2 knockdown in ISCs (Figure 6O). Considering the fact that additional three Gst genes are downregulated in the Wnt signaling-defective ISCs, these results indicate that Gst genes are required in ISCs to promote germ cell differentiation.

Both Cat and Gst genes work together to remove cellular hydrogen peroxide (H2O2). The c587-mediated Cat knockdown (CatKD) germaria behave like the control germaria, containing one CB on average (Figure 6M,O). Interestingly, c587-mediated CatKD significantly enhances the germ cell differentiation defect caused by GstD2KD1, and consequently the double knockdown germaria accumulate significantly more CBs than the single knockdown germaria (Figure 6N,O). In addition, the CatKD germaria carry a normal number of ISCs, whereas the GstD2KD germaria carry slightly fewer ISCs (Figure 6P,Q,S). Consistently, the GstD2KD1 CatKD germaria carry significantly fewer ISCs than the GstD1KD1 or CatKD germaria (Figure 6R,S). These results indicate that Cat and Gst genes work together in ISCs to maintain ISCs and promote germ cell differentiation, and further suggest that redox control is important for ISC maintenance.

Autocrine Wnt2 and Wnt4 function redundantly to maintain active Wnt signaling in ISCs

Our RNA-seq results indicate that four Wnt-like genes are expressed in the purified ISCs, but wingless shows little expression (Figure 7A). Wnt2 and Wnt4 are expressed at high levels, whereas Wnt5 and Wnt6 are present in much lower levels (Figure 7A). A recent study has shown that wnt4 mRNAs are indeed restricted to ISCs, whereas wnt2 mRNAs are also present in ISCs (Luo et al., 2015). Knocking down Wnt2 or Wnt4 alone in ISCs by two or three independent RNAi lines results in no or slight germ cell differentiation defect based on CB numbers (Figure 7B–D). This is in contrast with the recent study claiming that Wnt4 alone in ISCs is responsible for germ cell differentiation (Hamada-Kawaguchi et al., 2014). Consistent with the idea that Wnt2 and Wnt4 function redundantly in ISCs, simultaneous knockdown of Wnt2 and Wnt4 via different combinations of RNAi lines leads to a severe germ cell differentiation defect, which is similar to that caused by dshKD or armKD (Figure 7D–H). These results demonstrate that Wnt2 and Wnt4 in ISCs serve as redundant autocrine signals for promoting germ cell differentiation.

Wnt2 and Wnt4 function redundantly in ISCs to promote germ cell differentiation.

(A) RNA-seq results show that Wnt2 and Wnt4 are highly expressed in the purified ISCs in comparison with Wnt5 and Wnt6. In BC′ and EH, cap cells and GSCs are highlighted by ovals, CBs are indicated by arrowheads. (BD) Wnt2 (BB″) or Wnt4 (C, C′) knockdown by independent RNAi lines causes a slight accumulation of CBs. (D) Quantification results on CB numbers (the numbers on the top of bars represent the p values in comparison with the wild-type control, whereas those on the lines indicate the p values between single and double knockdown). (EH) Knocking down both Wnt2 and Wnt4 significantly increases CBs. (I) A schematic diagram showing that autocrine Wnt2 and Wnt4 signals control ISC maintenance and promote germ cell differentiation at least in part by maintaining the reduced redox state and preventing BMP signaling.

https://doi.org/10.7554/eLife.08174.013

Discussion

Although the differentiation niche is critical for promoting GSC progeny differentiation, little is known about its regulation. Here, we have identified Wnt2 and Wnt4 as autocrine signals to maintain the GSC differentiation niche partly through redox regulation (Figure 7I). First, Wnt signaling is required in ISCs for their maintenance by promoting cell proliferation and survival. Defective Wnt signaling causes a severe ISC loss, thereby preventing germ cell differentiation. Second, Wnt signaling is required to maintain the reduced redox state by sustaining the expression of Gst genes. This represents a novel connection between Wnt signaling and redox control. In addition, this study has also revealed that the reduced redox state is critical for ISC survival and thus for promoting germ cell differentiation. Third, Wnt signaling in ECs promotes GSC progeny differentiation partly by repressing BMP signaling in differentiated GSC progeny. Fourth, Wnt2 and Wnt4 represent redundant autocrine signals for maintaining Wnt signaling in the germ cell differentiation niche. Wnt signaling has been shown to control stem cell self-renewal directly (Holland et al., 2013), whereas ROS has shown to prime hematopoietic progenitor differentiation in Drosophila (Owusu-Ansah and Banerjee, 2009). This study has demonstrated the importance of Wnt signaling in maintaining the GSC differentiation niche by reducing ROS and thus promoting GSC lineage differentiation. Therefore, this study not only has identified critical signals for maintaining the GSC differentiation niche but also has revealed a novel function of Wnt signaling in the regulation of cellular redox.

This study has demonstrated that autocrine Wnt signaling controls ISC maintenance, thereby promoting germ cell differentiation. First, canonical Wnt signaling is required in ISCs to promote germ cell differentiation. c587-mediated knockdown of Wnt signal transducers Fz/fz2, dsh, and arm as well as c587-directed overexpression of Wnt signaling inhibitors sgg and axn leads to similar germ cell differentiation defects. Moreover, c587-directed overexpression of a constitutively active form of arm* can fully rescue the germ cell differentiation defect caused by dsh knockdown. Second, Wnt signaling maintains ISCs by promoting proliferation and survival. c587-mediated knockdown of dsh and arm leads to a severe ISC loss, whereas c587-mediated knockdown of sgg and axn or c587-directed arm* overexpression expands the ISC population. In addition, hyperactive Wnt signaling increases ISC proliferation and decreases apoptosis, whereas Wnt signaling downregulation increases ISC apoptosis and decreases proliferation. Thus, canonical Wnt signaling maintains the differentiation niche by promoting ISC proliferation and survival. However, we could not completely rule out the possibility that defective Wnt signaling leads to the loss of adult ISCs due to early developmental defects. Third, ISC-expressing Wnt2 and Wnt4 function redundantly in the differentiation niche to promote germ cell differentiation. Our RNA-seq results show that wnt2 and wnt4 mRNAs are present in the purified ISCs at high levels, while other wnt genes are expressed at much lower levels. c587-mediated wnt2 and wnt4 double knockdown results in more severe germ cell differentiation defects than wnt2 or wnt4 single knockdown.

Piwi has recently been shown to be required in the differentiation niche for promoting germ cell differentiation partly by repressing dpp expression (Jin et al., 2013; Ma et al., 2014). Although a recent study proposes that Wnt signaling controls germ cell differentiation by regulating piwi expression (Hamada-Kawaguchi et al., 2014), this study shows that Piwi protein and mRNAs are not downregulated in Wnt signaling-defective ISCs, suggesting that Wnt signaling does not sustain piwi expression in the differentiation niche to promote GSC progeny differentiation. Instead, this study has further revealed that Wnt signaling maintains the differentiation niche by controlling the cellular redox. First, our RNA-seq results show that GstD2, GstD4, GstD10, and GstE3 mRNA expression levels are significantly downregulated in the purified Wnt signaling-defective ISCs in comparison with the control ISCs. Second, defective Wnt signaling in ISCs elevates ROS levels in themselves and underneath germ cells, indicating that Wnt signaling is required in the differentiation niche to maintain low ROS levels. It remains unclear if increased ROS levels in early germ cells contribute to their differentiation defects. Third, c587-directed GST2, SOD1, and CAT overexpression can dramatically suppress the ROS elevation, the germ cell differentiation retardation, and the ISC loss caused by defective Wnt signaling, indicating that ROS elevation is responsible for the germ cell differentiation defect and the ISC loss. Finally, c587-mediated knockdown of GstD2 and Cat results in the ISC loss and the germ cell differentiation defect, indicating that ROS elevation in ISCs is sufficient to cause ISC loss and retard germ cell differentiation. This study has, for the first time, demonstrated that autocrine Wnt signaling controls cellular redox state in the differentiation niche and thus promotes germ cell differentiation.

So far, various studies have revealed a number of important signaling pathways and factors in ISCs to promote germ cell differentiation by maintaining ISC cellular process-mediated germ cell–soma interaction and preventing BMP signaling (Xie, 2013). This study shows that Wnt signaling is required for preventing BMP signaling in differentiated germ cells and for maintaining long ISC cellular processes. Rho, Eggless, Woc, Lsd1, Piwi, EGFR signaling, and ecdysone signaling have been shown to be required in ISCs to maintain long ISC cellular processes encasing germ cells (Schultz et al., 2002; Kirilly et al., 2011; Wang et al., 2011; Eliazer et al., 2014; Ma et al., 2014; Maimon et al., 2014; Konig and Shcherbata, 2015). Since properly differentiated germ cells are also required to maintain long ISC cellular processes (Kirilly et al., 2011), it is difficult to distinguish the cause and effect of the germ cell differentiation defect and the ISC cellular process loss. In contrast, three known strategies operate in the differentiation niche to prevent BMP signaling, thereby producing a permissive environment for germ cell differentiation. First, Lsd1, Rho, and Piwi are required in ISCs to repress dpp mRNA expression, thereby directly preventing BMP signaling in the differentiation niche (Eliazer et al., 2011; Kirilly et al., 2011; Jin et al., 2013; Ma et al., 2014). dpp encodes a BMP ligand, which activates BMP signaling important for GSC self-renewal (Xie and Spradling, 1998). Second, Rho, Eggless, and EGFR signaling function in ISCs to repress the expression of dally, preventing BMP diffusion from the self-renewal niche to the differentiation niche (Liu et al., 2010; Kirilly et al., 2011; Wang et al., 2011). Dally, which is a proteoglycan protein facilitating BMP differentiation and signaling, is expressed in cap cells, the GSC self-renewal niche, to restrict BMP signaling to GSCs (Akiyama et al., 2008; Guo and Wang, 2009). Third, a recent study has proposed that cap cell-initiated Wnt signaling maintains the expression of tkv encoding a type I BMP receptor in ISCs, thereby preventing BMP signaling in differentiated germ cells (Luo et al., 2015). Our findings from this study have also supported the idea that autocrine Wnt signaling in ISCs promotes GSC progeny differentiation partly by repressing BMP signaling. However, our findings have also argued that inactivating Wnt signaling does not regulate the mRNA expression levels of tkv and other BMP pathway components in ISCs. Therefore, future research will be needed to investigate the molecular mechanisms for Wnt signaling in the differentiation niche to prevent BMP signaling and maintain long ISC cellular processes.

Materials and methods

The Drosophila stocks used in this study include: c587 (Kirilly et al., 2011), PZ1444 (Kirilly et al., 2011), UAS-SOD1 (Pan et al., 2007), UAS-axn (BL7225), UAS-sgg (BL6818), armRNAi (BL31304; BL31305), dshRNAi (BL31306; BL31307), axnRNAi (BL31705), sggRNAi (BL35364), Wnt2RNAi (BL36722; TH00483; TH00484), and Wnt4RNAi (BL29442; TH00485). TH lines for Wnt2 and Wnt4 were constructed for this study according to the published procedure (Ni et al., 2011); the coding region of Gst2 and Cat were cloned into a UASp vector to generate UAS-Gst2 and UAS-Cat using standard molecular biology techniques. Drosophila strains were maintained and crossed at room temperature on standard cornmeal/molasses/agar media unless specified. To maximize the RNAi-mediated knockdown effect, newly eclosed flies were cultured at 29°C for a week before the analysis of ovarian phenotypes.

BrdU incorporation, TUNEL labeling, and DHE staining were performed according to the published procedures (Xie and Spradling, 1998; Owusu-Ansah and Banerjee, 2009; Wang et al., 2011). Immunohistochemistry was performed according to our previously published procedures (Xie and Spradling, 1998). The following antibodies were used in this study: rabbit polyclonal anti-β-galactosidase antibody (1:100, Cappel), mouse monoclonal anti-Hts antibody (1:50, DSHB), Guinea pig polyclonal anti-Piwi antibodies (1:100; provided by H. Lin), rabbit monoclonal anti-pS423/425 Smad3 antibody (1:100, Epitomics), and rat monoclonal anti-Vasa antibody (1:50, DSHB). All images were taken with a Leica TCS SP5 confocal microscope.

The GFP-positive dshKD or axn-overexpressing ISCs were sorted by FACS, and mRNA isolation was carried out according to the published procedure (Ma et al., 2014). Following manufacturer's directions and using the Illumina TruSeq Stranded mRNA LT Kit (Illumina, MA, United States; Cat. No. RS-122-2101/2), short fragment libraries were constructed. The resulting libraries were purified using Agencourt AMPure XP system (Backman Coulter, CA, United States; Cat. No. A63880) and were then quantified using a LabChip® GX (PerkinElmer, MA, United States) and a Qubit Fluorometer (Life Technologies, CA, United States). All libraries were pooled, re-quantified, and run as 50 bp single-end lanes on an Illumina HiSeq 2500 instrument, using HiSeq Control Software 2.0.5 and Real-Time Analysis (RTA) version 1.17.20.0. Secondary Analysis version CASAVA-1.8.2 was run to demultiplex reads and generate FASTQ files.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
    Shaggy and dishevelled exert opposite effects on Wingless and Decapentaplegic expression and on positional identity in imaginal discs
    1. TR Heslip
    2. H Theisen
    3. H Walker
    4. JL Marsh
    (1997)
    Development 124:1069–1078.
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
    The Drosophila fusome, a germline-specific organelle, contains membrane skeletal proteins and functions in cyst formation
    1. H Lin
    2. L Yue
    3. AC Spradling
    (1994)
    Development 120:947–956.
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
    Negative regulation of Armadillo, a Wingless effector in Drosophila
    1. LM Pai
    2. S Orsulic
    3. A Bejsovec
    4. M Peifer
    (1997)
    Development 124:2255–2266.
  35. 35
  36. 36
    Signaling from germ cells mediated by the rhomboid homolog stet organizes encapsulation by somatic support cells
    1. C Schultz
    2. CG Wood
    3. DL Jones
    4. SI Tazuke
    5. MT Fuller
    (2002)
    Development 129:4523–4534.
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
    A Drosophila axin homolog, daxin, inhibits Wnt signaling
    1. K Willert
    2. CY Logan
    3. A Arora
    4. M Fish
    5. R Nusse
    (1999)
    Development 126:4165–4173.
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46

Decision letter

  1. Utpal Banerjee
    Reviewing Editor; University of California, Los Angeles, United States

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

Thank you for submitting your work entitled “Wnt Signaling-Mediated Redox Regulation Maintains the Germline Stem Cell Differentiation Niche” for peer review at eLife. Your submission has been favorably evaluated by Fiona Watt (Senior Editor), a Reviewing Editor, and two reviewers.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

In this manuscript, Wang et al. address the role of somatic inner germarial sheath cells (ISCs) in promoting differentiation of germline cells displaced from the stem cell niche. Using a battery of genetic tools and molecular readouts, the Wnt signaling pathway is identified as a critical regulator of germ cell differentiation that acts, at least in part, autonomously within the ISCs. Clear evidence supports the conclusion that autonomous Wnt signaling is important for the viability of ISCs and that elevated ROS in Wnt signaling mutants contributes to the reduced viability of these cells. Conditions that lead to reduced ISC number led to accumulation of less differentiated germ cells, suggesting that Wnt-mediated maintenance of ISCs is important for controlling germ cell differentiation. The work adds a new mechanism to a developing field of study focused on understanding how individual signaling pathways that mediate interactions between ISC cells and germ cells influence germ cell biology.

Essential revisions:

The reviewers are unanimous in their appraisal of the manuscript and agree that if all the conclusions in the manuscript are properly supported by experiments, then this work will generate a really interesting new model for the signaling that controls somatic cell-germ cell interactions during germ cell differentiation. There are aspects of the analysis that are not conclusive, and these need to be fully addressed in order for a resubmission to be considered for publication.

The two central messages in this paper are:

1) Wnt2 and Wnt4 work redundantly and Wnt signaling in the ISC is somehow linked in a non-autonomous fashion to an IGS to germ cell signal, that contrary to previous reports is not piwi dependent.

2) The Wnt signal controls ROS signaling in the ISC cells and this is what provides the ISC to germ cell connection.

Of these two messages, the first one is a high-point of the analysis and we suggest some changes that will make the manuscript better. However, the equally important point 2 is not convincingly demonstrated. It is critical that these are addressed.

Point 1:

a) The role of Wnt4 in the process was previously known, but the redundancy with Wnt2 was not. On this point, rather than place the emphasis on what another laboratory “got wrong”, the authors could seize this opportunity to explain the general problem with redundancy and how this manuscript managed to avoid those pitfalls.

b) One assumes that the p-value resulting from loss of Wnt4 alone (using D1), in Figure 7D is not correct, certainly looks that way from the graph, but if somehow this p value is in the e-9 range (rather that about 0.1 or so), then we have a big problem with the redundancy argument.

c) The reviewers feel that the dismissal of the published piwi data needs some substantiation. While the authors' assurance that this is due to the use of whole ovaries not accounting for ISC loss seems reasonable, the authors did not discuss an alternative signal, and they need to address this issue more directly.

Genetic experiments demonstrating that piwi expression in ISCs (or possibly terminal filament and cap cells given the broad expression of the Gal4 driver utilized) can rescue Btk29 mutant defects in germ cell differentiation clearly indicate an important role for piwi in this process (Hamada-Kawaguchi). An explanation for the ability of piwi expression to rescue the effects of Btk29 mutation on germ cell differentiation and Bam expression should be considered.

In addition, the possibility that loss of interactions between ISC projections and differentiated germ cells (described in Kirilly, et al. 2011) upon dissociation of the germarium tissue for FACS sorting might alter the in vivo effects of genetic mutations on piwi expression should be carefully considered.

Finally, the armKD image shown in Figure 5B and B' appears to have reduced piwi expression in many of the ISC cells, and the methods for quantifying Piwi protein levels are not clear. Rigorous exclusion of piwi as relevant for Wnt-mediated germ cell differentiation regulation would require an experiment in which piwi is expressed in ISCs lacking arm or dsh and assessing its ability to rescue the defects observed. The main point is that genetic evidence supports roles within ISCs for all of the above genes, and the manuscript would be enhanced by the proposal of an integrated model that considers the individual roles of each pathway in this process and how these coordinate to control germ cell differentiation.

In this context, the authors should comment on how the entire network of genes such Btk29, EGFR, Rho, Piwi, and BMP that are key factors for controlling the ISC-germ cell interactions that promote germ cell differentiation, integrates into the Wnt model. This can be addressed by textual changes only if the authors think this will enhance the paper.

Some of the Gal4 drivers used in this study are not cell type specific. It will be important to use Cap and Follicle cell drivers to confirm that the Wnt effects as stated are indeed specific to the ISC. This is an important caveat to address.

Point 2:

The ROS connection to Wnt signaling needs to be addressed with better data.

a) It is not clear, from the DHE staining shown, that knockdown of dsh affects ROS levels in ISCs. For example, the difference in DHE staining in 6B vs 6B' could be because the ISC cells are absent in a dshKD context (as suggested elsewhere in the paper). So this does not definitively establish that loss of Wnt signaling affects ROS levels in ISC cells.

b) The RNAseq data are supportive and correlative, but on its own, it is not enough to establish causality. To test whether the decrease in scavengers is the cause of the dshKD phenotype, it would be important to determine whether knockdown of scavengers also causes a loss of ISCs, as knockdown of dsh does.

c) The proliferation and apoptosis phenotypes described in Figure 3 may be due to effects during development, which would call into question whether Wnt promotes survival and proliferation during adulthood.

Here are some suggestions from the reviewers, but if the authors think of better experiments to substitute for these that will convince the reviewers, then this will be OK:

i) Controls with scavenger knockdown alone to show the DHE staining works as expected, and counter staining (e.g. LacZ in PZ1444) to identify IGS cells and show at cellular resolution the level of DHE staining in wt vs dshKD conditions. Since it seems all the ISC cells are gone by 7 days after heat shift, it may be necessary to either look earlier, before all the ISC cells have disappeared, or prevent apoptosis by some other means, such as overexpression of p35. If they find that DHE staining is actually higher in dshKD ISC cells vs wt, this would go a long way toward addressing the reviewers' concerns.

ii) Quantification of ISC number in GSTD2 + Cat double KD will determine if the ISC number changes or remains the same. The result needs to fit the model.

iii) To determine whether the proliferation and apoptosis phenotypes described in Figure 3 are due to effects during development, the authors should use tsGal80 to suppress expression until adulthood, as in Figure 2, and perform TUNEL staining of dshKD ISC cells as they did for arm*.

iv) Does germ-cell lethality revert when ROS scavengers are expressed in them?

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled “Wnt Signaling-Mediated Redox Regulation Maintains the Germline Stem Cell Differentiation Niche” for further consideration at eLife. Your revised article has been favorably evaluated by Fiona Watt (Senior Editor), a Reviewing Editor, and two reviewers. The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below:

1) An integrated model of how Wnt signaling and BMP signaling might work together is confusing, more so in this version of the manuscript than the last. pMAD and Dad-lacZ are dramatically elevated in undifferentiated germ cells in germaria with reduced dsh or arm activity (Figure 4), but the final conclusion is that BMP signaling is not affected by altered Wnt because no changes in transcript levels of other dpp effectors (including dad?) were detected and because a 50% reduction of dpp ligand did not alter the observed phenotypes, except in the case of the dshKD1 background, where it appears to have rescued the phenotype quite dramatically. Given these data, the conclusion about independence of the pathways is not justified and needs to be tempered and explained better.

2) ROS induced in the germ cells acts downstream of the IGS cell Wnt signal to inhibit differentiation. Is there a rescue of normal differentiation by reducing ROS in germ cells when IGS cells still have reduced Wnt signaling? (A definitive comment on the link of Wnt signaling, ROS and differentiation will make this paper very exciting. However, if tools to test this conclusion are time consuming to put together, then adding this as a point of discussion will suffice).

3) One of our reviewers expressed concern that Wnt signaling during development might produce defective adult tissues in which proliferation and survival of ISCs is different. This can be easily addressed by repeating the proliferation and apoptosis experiments with arm*, armKD, and dshKD (Figure 3) with Gal80ts. Please pick at least one (or more) genotype to test with Gal80 to demonstrate that this is not the case. If this will involve multi generational crosses and a significant delay in revision, then please bring this up as an important caveat for the whole study.

4) Given the large effect of Wnt4 knockdown, the redundancy argument is not strong. Also, the data should include p values for should provide p-values for (Wnt2 vs Wnt2+4 and Wnt4 vs Wnt2+4.

5) c587 is described in all sections of the paper as "ISC-specific”. This is not true and needs to be changed.

6) The new data on genetic alteration of ROS levels focused on germaria with 2 cystoblasts as indicative of a differentiation phenotype. In the text associated with Figures 1 and 2, it states that most germaria have 1-2 cystoblasts, and only gemaria with more than 4 Hts/spectrosome positive germaria were scored. It is not clear how only 2 cystoblasts can be accurately scored as a "differentiation defect” if most germaria have 1 or 2 cystoblasts present. Most likely, this is due to redundancy among GST genes, but it's hard to know how to compare this data to that presented in Figure 1.

https://doi.org/10.7554/eLife.08174.014

Author response

Point 1:

a) The role of Wnt4 in the process was previously known, but the redundancy with Wnt2 was not. On this point, rather than place the emphasis on what another laboratory “got wrong”, the authors could seize this opportunity to explain the general problem with redundancy and how this manuscript managed to avoid those pitfalls.

Thank the reviewers for the comment. In the revised manuscript, we have put the emphasis on the requirement of both Wnt2 and Wnt4 in ISCs for maintaining the differentiation niche and promoting germ cell differentiation.

b) One assumes that the p-value resulting from loss of Wnt4 alone (using D1), in Figure 7D is not correct, certainly looks that way from the graph, but if somehow this p value is in the e-9 range (rather that about 0.1 or so), then we have a big problem with the redundancy argument.

Indeed, knocking down wnt2 or wnt4 by one RNAi line yields significantly more CBs than the controls, indicating that both Wnt2 and Wnt4 are required for sustaining maximal Wnt signaling in ISCs to promote germ cell differentiation. However, simultaneous knockdown of both wnt2 and wnt4 in ISCs produces much severe germ cell differentiation defects, which is more than the additive effect of two single knockdowns. These results strongly argue that Wnt2 and Wnt4 function redundantly in ISCs to promote germ cell differentiation.

c) The reviewers feel that the dismissal of the published piwi data needs some substantiation. While the authors' assurance that this is due to the use of whole ovaries not accounting for ISC loss seems reasonable, the authors did not discuss an alternative signal, and they need to address this issue more directly.

Genetic experiments demonstrating that piwi expression in ISCs (or possibly terminal filament and cap cells given the broad expression of the Gal4 driver utilized) can rescue Btk29 mutant defects in germ cell differentiation clearly indicate an important role for piwi in this process (Hamada-Kawaguchi). An explanation for the ability of piwi expression to rescue the effects of Btk29 mutation on germ cell differentiation and Bam expression should be considered.

This reviewer might have misunderstood our argument. Our results on piwi mRNA and protein have clearly demonstrated that piwi expression is not downregulated in Wnt signaling-defective ISCs as suggested by Hamada-Kawaguchi et al. Others and we have confirmed the important function of Piwi in ISCs to promote germ cell differentiation (Jin et al., 2013; Ma et al. 2014). Even though Piwi overexpression can “bypass” the requirement of Wnt signaling in ISCs in promoting germ cell differentiation, it does not necessarily support the Hamada-Kawaguchi model that Wnt signaling controls the expression of Piwi to promote germ cell differentiation. However, the rescue experimental results do suggest that Piwi might function downstream (though not directly regulated by Wnt signaling) or in parallel with Wnt signaling in ISCs to promote germ cell differentiation. We have re-written the part to avoid any potential confusion.

In addition, the possibility that loss of interactions between ISC projections and differentiated germ cells (described in Kirilly, et al. 2011) upon dissociation of the germarium tissue for FACS sorting might alter the in vivo effects of genetic mutations on piwi expression should be carefully considered.

We thank the reviewers for the interesting thought. The dissociation process is rapid, and the cell-sorting process is carried out at 4°C, which prevents transcriptional changes. Although we could not completely ruled out the suggested possibility, it is very unlikely because we have shown that Piwi protein levels are similar or upregulated in armKD/dshKD ISCs in INTACT germaria. Based on our results on piwi mRNA and protein, we are confident that defective Wnt signaling does not lead to the downregulated piwi expression in ISCs.

Finally, the armKD image shown in Figure 5B and B' appears to have reduced piwi expression in many of the ISC cells, and the methods for quantifying Piwi protein levels are not clear. Rigorous exclusion of piwi as relevant for Wnt-mediated germ cell differentiation regulation would require an experiment in which piwi is expressed in ISCs lacking arm or dsh and assessing its ability to rescue the defects observed. The main point is that genetic evidence supports roles within ISCs for all of the above genes, and the manuscript would be enhanced by the proposal of an integrated model that considers the individual roles of each pathway in this process and how these coordinate to control germ cell differentiation.

For Piwi protein expression quantification, we used the Leica quantification software to select the nucleus area and measure the total fluorescence intensity. Like any experiments, there are some degrees of variations due to expression variation, staining variation or both. After measuring 203 wild-type ISCs, 106 armKD ISCs and 139 dshKD ISCs, we could not find any evidence supporting that Piwi protein levels are down-regulated in Wnt signaling-defective ISCs. These findings on piwi mRNA and protein levels are mutually supported. In this study, we argue that Piwi is NOT the direct target of Wnt signaling in ISCs. We also believe that Piwi and Wnt signaling are relevant with each other in controlling ISC maintenance and promoting germ cell differentiation because they have similar mutant phenotypes in ISC maintenance and germ cell differentiation.

In this context, the authors should comment on how the entire network of genes such Btk29, EGFR, Rho, Piwi, and BMP that are key factors for controlling the ISC-germ cell interactions that promote germ cell differentiation, integrates into the Wnt model. This can be addressed by textual changes only if the authors think this will enhance the paper.

According to the suggestion, we have done our best to add some of our thoughts on the relationships among the pathways in the Discussion though the connections are largely unknown. Hopefully, these speculations will help stimulate future research in these areas.

Some of the Gal4 drivers used in this study are not cell type specific. It will be important to use Cap and Follicle cell drivers to confirm that the Wnt effects as stated are indeed specific to the ISC. This is an important caveat to address.

As the reviewers pointed out, c587 expresses GAL4 in not only in ISCs but also in early follicle precursor cells. In addition, the existing Gal4 drivers for cap cells, bab1-gal4 and hh-gal4, are also expressed in ISCs at low levels. There is no good GAL4 driver, which is specifically expressed in early follicle progenitors. Therefore, it is very challenging to use various GAL4 drivers to address the caveat. A recently published study by the Yu Cai group shows that wnt4 mRNAs are primarily present in ISCs, and wnt2 mRNAs are also expressed in ISCs. These results are consistent with our RNA-seq and genetic knockdown results.

Point 2:

[…] Here are some suggestions from the reviewers, but if the authors think of better experiments to substitute for these that will convince the reviewers, then this will be OK:

i) Controls with scavenger knockdown alone to show the DHE staining works as expected, and counter staining (e.g. LacZ in PZ1444) to identify IGS cells and show at cellular resolution the level of DHE staining in wt vs dshKD conditions. Since it seems all the ISC cells are gone by 7 days after heat shift, it may be necessary to either look earlier, before all the ISC cells have disappeared, or prevent apoptosis by some other means, such as overexpression of p35. If they find that DHE staining is actually higher in dshKD ISC cells vs wt, this would go a long way toward addressing the reviewers' concerns.

According to the request, we have performed DHE staining in PZ1444-labeled armKD and dshKD germaria before most of the ISCs are lost. Our results confirm that dshKD or armKD ISCs exhibit elevated levels of DHE staining. New results have been added to Figure 6—figure supplement 1B–F′.

ii) Quantification of ISC number in GSTD2 + Cat double KD will determine if the ISC number changes or remains the same. The result needs to fit the model.

According to the request, we have quantified ISC numbers in single and double knockdown germaria for GSTD2 and Cat. Indeed, the ISC number is significantly reduced in GSTD2 Cat double knockdown germaria, which correlates with the germ cell differentiation defect. The new results have been added to Figure 6P-S.

iii) To determine whether the proliferation and apoptosis phenotypes described in Figure 3 are due to effects during development, the authors should use tsGal80 to suppress expression until adulthood, as in Figure 2, and perform TUNEL staining of dshKD ISC cells as they did for arm*.

According to the request, we have performed both BrdU and TUNEL experiments on the dshKD and armKD germaria. Our results show that knocking down dsh or arm indeed leads to significantly decreased BrdU-positive ISCs and significantly increased TUNEL-positive ISCs. Now new results have been provided in Figure 3H–Q.

iv) Does germ-cell lethality revert when ROS scavengers are expressed in them?

When arm and dsh are knocked down in ISCs, the severe germ cell differentiation defect appears, but the GSC loss is rarely observed. When ROS scavengers, including Cat, SOD and GST2, are expressed in ISCs, the germ cell differentiation defect caused by arm and dsh knockdown is drastically rescued, but GSC numbers remain largely unchanged. We have added the new results in Figure 6—figure supplement 2G.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

1) An integrated model of how Wnt signaling and BMP signaling might work together is confusing, more so in this version of the manuscript than the last. pMAD and Dad-lacZ are dramatically elevated in undifferentiated germ cells in germaria with reduced dsh or arm activity (Figure 4), but the final conclusion is that BMP signaling is not affected by altered Wnt because no changes in transcript levels of other dpp effectors (including dad?) were detected and because a 50% reduction of dpp ligand did not alter the observed phenotypes, except in the case of the dshKD1 background, where it appears to have rescued the phenotype quite dramatically. Given these data, the conclusion about independence of the pathways is not justified and needs to be tempered and explained better.

The reviewers must have misread our manuscript. We show that some accumulated CB-like cells are positive for pMad and Dad-lacZ expression, indicative of upregulated BMP signaling activities. Our quantitative results in Figure 4H–H′ show that the heterozygous dpp mutations, dpphr4 and dpphr56, can partially rescue the germ cell differentiation defect caused by dshKD and dshKD2, indicating “Wnt signaling in ISCs regulates germ cell differentiation partly by preventing BMP signaling in the differentiation niche.” However, we show that Wnt signaling does not regulate mRNA levels for BMP pathway components in ISCs, which might give the reviewers the impression “BMP signaling is not affected by altered Wnt (signaling)”. In the revised manuscript, we have made it clear that Wnt signaling in ISCs is required to prevent BMP signaling, and add the information to the model in Figure 7I.

2) ROS induced in the germ cells acts downstream of the IGS cell Wnt signal to inhibit differentiation. Is there a rescue of normal differentiation by reducing ROS in germ cells when IGS cells still have reduced Wnt signaling? (A definitive comment on the link of Wnt signaling, ROS and differentiation will make this paper very exciting. However, if tools to test this conclusion are time consuming to put together, then adding this as a point of discussion will suffice).

That is a great question. All the existing tools cannot answer the question satisfactorily. We are in the process of generating new transgenic fly strains, allowing us to conditionally remove ROS in germ cells while Wnt signaling are still defective in IGS cells. It will take a long time to generate new transgenic fly strains and bring all the genetic elements together to accomplish this goal. We will leave this question for the future studies. In the Discussion, we have added the sentence “In the future, it will be important to determine if increased ROS levels in early germ cells contribute to their differentiation defects caused by defective Wnt signaling in ISCs”.

3) One of our reviewers expressed concern that Wnt signaling during development might produce defective adult tissues in which proliferation and survival of ISCs is different. This can be easily addressed by repeating the proliferation and apoptosis experiments with arm*, armKD, and dshKD (Figure 3) with Gal80ts. Please pick at least one (or more) genotype to test with Gal80 to demonstrate that this is not the case. If this will involve multi generational crosses and a significant delay in revision, then please bring this up as an important caveat for the whole study.

In order to address this question, we would need at least a few months to bring five transgenes together, which also requires chromosomal recombination. Based on our existing experimental results, this is NOT a big issue. Our experimental results have shown that Wnt signaling is required in adult ISCs to promote germ cell differentiation and maintain themselves. In addition, we have gain-of-function and loss-of-function experiments to demonstrate that Wnt signaling is required to maintain ISCs by regulating proliferation and survival. Since we are not able to provide the experimental results excluding the possibility, we have added the statement “However, we could not completely rule out the possibility that defective Wnt signaling leads to the loss of adult ISCs due to their developmental defects.”

4) Given the large effect of Wnt4 knockdown, the redundancy argument is not strong. Also, the data should include p values for should provide p-values for (Wnt2 vs Wnt2+4 and Wnt4 vs Wnt2+4.

As suggested, we have added the P values for Wnt2 versus Wnt2+4 and Wnt4 versus Wnt2+4 in Figure 7D. Our results have indicated that Wnt2 and 4 double knockdowns have significantly severe germ cell differentiation defects than Wnt2 and Wnt4 single knockdowns, which support the redundant roles of Wnt2 and Wnt4.

5) c587 is described in all sections of the paper as “ISC-specific”. This is not true and needs to be changed.

As suggested, we have modified the “ISC-specific” into “c587-mediated”.

6) The new data on genetic alteration of ROS levels focused on germaria with 2 cystoblasts as indicative of a differentiation phenotype. In the text associated with Figures 1 and 2, it states that most germaria have 1-2 cystoblasts, and only gemaria with more than 4 Hts/spectrosome positive germaria were scored. It is not clear how only 2 cystoblasts can be accurately scored as a “differentiation defect” if most germaria have 1 or 2 cystoblasts present. Most likely, this is due to redundancy among GST genes, but it's hard to know how to compare this data to that presented in Figure 1.

A normal germarium contains 0 to 2 CBs with an average of one CB. Due to the weak differentiation defect of GstD2KD+CatKD, we have quantified the exact CB number to determine if there is any germ cell differentiation defect. Our statistic analysis results allow us to conclude that there is any weak and yet significant differentiation defect. In Figure 4H–H′, dshKD1 and dshKD2 germaria contain 15 and 13 CBs, respectively. Using the germaria carrying 4 or more CBs as the differentiation defect is a more efficient quantification method in Figures 1 and 2.

https://doi.org/10.7554/eLife.08174.015

Article and author information

Author details

  1. Su Wang

    1. Stowers Institute for Medical Research, Kansas City, United States
    2. Department of Anatomy and Cell Biology, University of Kansas School of Medicine, Kansas City, United States
    Contribution
    SW, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    The authors declare that no competing interests exist.
  2. Yuan Gao

    Center for Life Sciences, College of Life Sciences, School of Medical Sciences, Tsinghua University, Beijing, China
    Contribution
    YG, Acquisition of data, Analysis and interpretation of data
    Competing interests
    The authors declare that no competing interests exist.
  3. Xiaoqing Song

    Stowers Institute for Medical Research, Kansas City, United States
    Contribution
    XS, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  4. Xing Ma

    1. Stowers Institute for Medical Research, Kansas City, United States
    2. Department of Anatomy and Cell Biology, University of Kansas School of Medicine, Kansas City, United States
    Contribution
    XM, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  5. Xiujuan Zhu

    Stowers Institute for Medical Research, Kansas City, United States
    Contribution
    XZ, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  6. Ying Mao

    Center for Life Sciences, College of Life Sciences, School of Medical Sciences, Tsinghua University, Beijing, China
    Contribution
    YM, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  7. Zhihao Yang

    Center for Life Sciences, College of Life Sciences, School of Medical Sciences, Tsinghua University, Beijing, China
    Contribution
    ZY, Contributed unpublished essential data or reagent
    Competing interests
    The authors declare that no competing interests exist.
  8. Jianquan Ni

    Center for Life Sciences, College of Life Sciences, School of Medical Sciences, Tsinghua University, Beijing, China
    Contribution
    JN, Contributed unpublished essential data or reagent
    Competing interests
    The authors declare that no competing interests exist.
  9. Hua Li

    Stowers Institute for Medical Research, Kansas City, United States
    Contribution
    HL, Analysis and interpretation of data
    Competing interests
    The authors declare that no competing interests exist.
  10. Kathryn E Malanowski

    Stowers Institute for Medical Research, Kansas City, United States
    Contribution
    KEM, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  11. Perera Anoja

    Stowers Institute for Medical Research, Kansas City, United States
    Contribution
    PA, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  12. Jungeun Park

    Stowers Institute for Medical Research, Kansas City, United States
    Contribution
    JP, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  13. Jeff Haug

    Stowers Institute for Medical Research, Kansas City, United States
    Contribution
    JH, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  14. Ting Xie

    1. Stowers Institute for Medical Research, Kansas City, United States
    2. Department of Anatomy and Cell Biology, University of Kansas School of Medicine, Kansas City, United States
    Contribution
    TX, Conception and design, Analysis and interpretation of data, Drafting or revising the article
    For correspondence
    tgx@stowers.org
    Competing interests
    The authors declare that no competing interests exist.

Funding

Stowers Institute for Medical Research (The SIMR Fund)

  • Ting Xie

The funder had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to thank Dr H Lin, Developmental Studies Hybridoma Bank and Bloomington Drosophila Stock Center for reagents; the Xie laboratory members for stimulating discussions; Drs N Sachan and D Chao for critical comments on the manuscript. This work was supported by the Stowers Institute for Medical Research (TX).

Reviewing Editor

  1. Utpal Banerjee, University of California, Los Angeles, United States

Publication history

  1. Received: April 17, 2015
  2. Accepted: September 12, 2015
  3. Version of Record published: October 9, 2015 (version 1)

Copyright

© 2015, Wang et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,565
    Page views
  • 397
    Downloads
  • 19
    Citations

Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)