Btk29A Promotes Wnt4 Signaling in the Niche to Terminate Germ Cell Proliferation in Drosophila

See allHide authors and affiliations

Science  17 Jan 2014:
Vol. 343, Issue 6168, pp. 294-297
DOI: 10.1126/science.1244512

Wnt–β-Catenin in Germ Cells

The Wnt–β-catenin pathway contributes to many signaling mechanisms during organismal development and carcinogenesis by regulating both transcription and cell adhesion. Hamada-Kawaguchi et al. (p. 294) demonstrate that this pathway must be activated in ovarian somatic cells to stop proliferation of germ cells in Drosophila. Phosphorylation of a tyrosine residue on β-catenin by the tyrosine kinase Btk turns on signaling in the niche cells by promoting transcriptional activity of β-catenin. Failure in this process resulted in ovarian tumors in the flies.


Btk29A is the Drosophila ortholog of the mammalian Bruton’s tyrosine kinase (Btk), mutations of which in humans cause a heritable immunodeficiency disease. Btk29A mutations stabilized the proliferating cystoblast fate, leading to an ovarian tumor. This phenotype was rescued by overexpression of wild-type Btk29A and phenocopied by the interference of Wnt4–β-catenin signaling or its putative downstream nuclear protein Piwi in somatic escort cells. Btk29A and mammalian Btk directly phosphorylated tyrosine residues of β-catenin, leading to the up-regulation of its transcriptional activity. Thus, we identify a transcriptional switch involving the kinase Btk29A/Btk and its phosphorylation target, β-catenin, which functions downstream of Wnt4 in escort cells to terminate Drosophila germ cell proliferation through up-regulation of piwi expression. This signaling mechanism likely represents a versatile developmental switch.

Stem cell maintenance and differentiation are not entirely autonomic, but instead are under strict control by supporting cells that form the “niche” (fig. S1A) (1, 2). Recent studies in Drosophila have shown that the dynamics of Piwi and its associated piRNAs, a protein-RNA complex for gene silencing, are required in not only germ cells but also distinct niche-forming somatic cells—escort cells—for germ cell development (3, 4); however, their regulatory mechanisms remain largely unknown. Here we identify a transcriptional switch involving the factor Bruton’s tyrosine kinase (Btk) and its phosphorylation target, β-catenin, operating downstream of Wnt4 in escort cells to terminate Drosophila germ cell proliferation through modulation of piwi expression.

Drosophila Btk29A type 2 is the ortholog of human BTK (fig. S1B) (5). The type 1 isoform is present and the type 2 is absent in Btk29AficP mutants (fig. S1G) (5). Germ stem cells (GSCs) and transit amplifying cystoblasts (CBs) are localized in the germarium situated at the anterior tip of an ovariole, posteriorly flanked by region 2, in which each CB divides twice and differentiates into cystocytes. The 16 cystocytes originating from a single CB remain interconnected by the fibrous structure fusome, a derivative of the spectrosome. GSCs and CBs both carry the spectrosome, a round, tubulin-enriched structure. The Btk29A mutant germarium contains significantly more germ cells than does the wild-type germanium (Fig. 1, A to J, and fig. S1, C to F). Although we observed supernumerary cells with spectrosomes in the Btk29AficP germarium (Fig. 1, A to K), many of the excess cells appear to be cystocytes, as they were accompanied by a branched fusome structure (Fig. 1E). A large excess of cystocytes in grossly deformed ovarioles has been observed in female Drosophila that are mutant for mei-P26 (6), a gene encoding a TRIM-NHL (tripartite motif and Ncl-1, HT2A, and Lin-41 domain) protein that binds to the argonaute protein Ago-1 for microRNA regulation. In mei-P26 mutants, an ovarian tumor “cystocytoma” is formed because cystocytes regain the ability to self-renew after they enter the differentiation path (6). This suggests that mei-P26 normally terminates CB proliferation. Intriguingly, the following phenotypes of mei-P26 were recapitulated in Btk29AficP. First, phospho-histone H3–positive mitotic germline cells, which were restricted to the anterior tip of the wild-type germarium (fig. S1O), were detected throughout the ovarioles (fig. S1P). Second, the expression of Bam, a protein that induces differentiation of GSCs into CBs in the wild type, was markedly increased in CB-like GSC daughters (fig. S1, Q and R). Third, oo18 RNA-binding protein (Orb) remained expressed in multiple cells in a cyst (fig. S1, S and T), contrasting to a wild-type cyst, where Orb expression becomes restricted to an oocyte.

Fig. 1 Deficiency of Btk29A in niche cells impairs germ cell development.

(A to I) The germaria of flies of the indicated genotypes doubly stained with anti-Vasa and monoclonal antibody (mAb) 1B1. Cells with spectrosomes observed in (B), (E), and (H) are schematically drawn in (C), (F), and (I). (J and K) The number of germ cells (J) and spectrosomes (K) in the region anterior to region 2 (indicated in fig. S1A, bottom) of the germaria of the indicated fly genotypes. In (J), the number of germ cells was estimated by counting the number of Vasa-expressing cells present on the single optical section that gave the largest surface area among all horizontal sagittal sections from a germarium. In (K), the numbers of spectrosomes contained in a stack image of a whole germarium are shown. Values are presented as the mean ± SEM. The number of germaria examined is indicated in parentheses. Statistical differences were evaluated by Student’s t test (**P < 0.01).

The reduction in mei-P26 transcription in Btk29AficP (Fig. 2, A and B) places mei-P26 downstream of Btk29A. Notably, mei-P26 functions cell-autonomously in germ cells (6, 7). However, the almost complete rescue of germ cell defects in Btk29AficP was attained by overexpression of Btk29A+ type 2 via bab1-Gal4 (Fig. 1, G to K), which showed high levels of expression in terminal filament cells and cap cells (TF and CPC in fig. S1A, respectively) and lower levels of expression in escort cells (EC in fig. S1A). bab1-Gal4 was effective in inducing germ cell overproduction when used to knockdown Btk29A (Fig. 2I and fig. S1, J and K). We also used hh-Gal4 with expression in the terminal filament cells and cap cells and c587-Gal4 with expression in escort cells to target UAS-Btk29ARNAi expression; c587-Gal4, but not hh-Gal4, led to the overproduction of spectrosome-bearing cells (Fig. 2, C to F and I, and fig. S1, L and M), and therefore, the escort cells were considered as likely sites of Btk29A action. These observations imply that Btk29A is required in the escort cells for soma-to-germ signaling to control the switch from proliferation to differentiation in germ cells, where mei-P26 functions as a core component of the switch (fig. S6).

Fig. 2 Effect of somatic knockdown of Btk29A or piwi.

(A and B) Reverse transcriptase–polymerase chain reaction analysis of piwi, dpp, mei-P26, and gbb. actin served as a control. The number of ovaries used for the analysis was standardized to yield equivalent actin signals in both the wild-type flies and mutants. (C to H) Germaria of flies of the indicated genotypes. The flies with Btk29A type 2 RNAi (RNA interference) carried also a copy of UAS-Dcr2. (I) The number of spectrosomes of flies of the indicated genotypes. Values represent mean ± SEM (**P < 0.01, Student’s t test); ns, not significant. (J) Bam expression in germ cells was increased by somatic knockdown of piwi. (K and L) Localization of Piwi (K) or Btk29A type 2 (L) as revealed by immunostaining. Scale bar: 10 μm.

Bone morphogenetic protein (BMP) signaling and piwi-dependent signaling compose two different pathways in the niche to control proliferation and differentiation of GSCs and their daughters (1, 8, 9). BMPs are secreted morphogens, and Piwi is an argonaute protein regulating gene expression. The Btk29AficP mutation abrogated piwi expression with little effect on decapentaplegic (dpp) or glass bottom boat (gbb) expression, two BMPs operating in the germarium (Fig. 2, A and B), and the BMP downstream component Mothers against Dpp (Mad) was normally phosphorylated in Btk29AficP GSCs (fig. S1, H and I). Furthermore, somatic piwi knockdown mimicked the Btk29AficP ovarian phenotypes (Fig. 2, G to J, and fig. S2, E to J).

Immunohistochemistry revealed that the Btk29AficP mutation or somatic Btk29A knockdown abrogated Piwi expression in the niche, but not in germ cells (Fig. 2K; see also fig. S1N). This reduction in Piwi expression was reversed by the somatic Btk29A+ overexpression (Fig. 2K). Furthermore, the loss-of-function piwi allele dominantly enhanced the Btk29A mutant phenotype (Fig. 3A and fig. S2, K to T). Moreover, somatic overexpression of piwi+ in Btk29AficP alleviated the germ cell hypertrophy (fig. S2, U and V) and reduced Bam expression to the normal level (Fig. 3, B to D). We therefore consider that Btk29A regulates the Piwi-dependent pathway in the niche to control germ cell proliferation.

Fig. 3 Genetic interactions between Btk29A and piwi and tyrosine phosphorylation of Arm by Btk29A.

(A) The number of spectrosomes in the indicated fly groups. Values represent mean + SEM (**P < 0.01, Student’s t test); ns, not significant. (B to D) Bam protein expression. (E to G, N) Western analysis of the immunoprecipitates of ovarian lysates with antibodies (indicated as IP) was probed with the same or other antibodies (indicated as Blot) (19). β-Tubulin served as a loading control. (G) In vitro phosphorylation (19) of Arm immunoprecipitated from Btk29AficP. mAB-Btk29A(COM), which recognizes both the type 1 and type 2 isoforms of Btk29A, was used in immunoprecipitation. (H to M) Anti-Arm staining. Escort cells are indicated by arrows. Scale bar: 10 μm. [(J) and (M) are schematic drawings of the germarium; TF: terminal filament cell, CPC: cap cell, GSC: germ stem cell, CB: cystoblast, EC: escort cell.] (N) The amount of Arm coimmunoprecipitated with DE-cadherin was greater in Btk29AficP than in the wild type.

Piwi and piRNAs constitute a major transposon-silencing pathway (1012). Somatic knockdown of Btk29A resulted in an increase in the expression of gypsy-lacZ (13) that monitored the activity of the gypsy transposon (fig. S3, A and B). Also, transcript levels of the ZAM, DM412, and mdg1 transposons were significantly increased in Btk29AficP (fig. S3C). We conclude that the Piwi deficiency due to the impairment of Btk29A results in derepression of transposon activities.

Genome instability associated with transposon mobilization may lead to the activation of a DNA double-strand break (DSB) checkpoint (14, 15). We found that a mutation in DSB signaling, mnk, did not ameliorate the germ cell phenotype induced by somatic Btk29A knockdown (fig. S3, D to H), indicating that the germ cell hypertrophy by the Btk29A deficiency is not a consequence of the DSB checkpoint activation.

Next, we searched for potential substrates of Btk29A in the niche. Btk29A type 2 was enriched in the interface between cells (Fig. 2L), where Drosophila melanogaster epithelial (DE)–cadherin and associated Arm, the β-catenin ortholog, are the major structural components (16). We found no sign of tyrosine phosphorylation of DE-cadherin (fig. S4, F and G), whereas Arm contained a high level of phosphotyrosine (Fig. 3, E and F), which was almost entirely absent from Btk29AficP ovaries (Fig. 3E). However, Arm immunoprecipitated from Btk29AficP was strongly phosphorylated in vitro by the exposure of Arm to active Btk29A protein that had been immunoprecipitated from wild-type ovaries (Fig. 3G). These results demonstrate that Btk29A mediates the tyrosine phosphorylation of Arm in vivo.

The anti-Arm labeling intensity of cell adhesion sites was stronger in Btk29AficP (Fig. 3, K to M) than in the wild type (Fig. 3, H to J). Immunoprecipitation assays revealed that the relative amount of Arm associated with DE-cadherin was greater in Btk29AficP than in the wild type (Fig. 3N), suggesting that the tyrosine phosphorylation of Arm facilitates its release from the membrane to the cytoplasm, as in mammalian cells (17).

Mammalian β-catenin is tyrosine-phosphorylated at residues Y86, Y142, and Y654 (fig. S4A). When transfected into mammalian Cos7 cells, Drosophila Btk29A type 2 phosphorylated all these tyrosine residues of β-catenin (fig. S4, B to D). Moreover, the antibodies against phosphorylated Y142 (anti-pY142) and anti-pY654 recognized Arm phosphorylated at the conserved site Y150 and Y667, respectively, in the immunoprecipitates from ovarian lysates (Fig. 4A).

Fig. 4 Btk29A/Btk-dependent tyrosine phosphorylation of Arm/β-catenin.

(A) Western blotting of anti-Arm immunoprecipitates of ovarian lysates probed with anti-pY142 or anti-pY654. (B to I) Germaria of flies of the indicated genotypes. (F to I) Piwi expression (F and H) or Bam expression (G and I) in germaria of flies of the indicated genotypes. (J) The number of spectrosomes. Values represent mean + SEM (**P < 0.01, Student’s t test); ns, not significant.

Expression of unphosphorylatable Arm-Y150F in the escort cells via c587-GAL4 or bab1-GAL4, but not hh-Gal4, induced germ cell hypertrophy (fig. S5I), whereas another unphosphorylatable mutant, Y667F, or wild-type Arm exerted little effect (Fig. 4, B to E and J, and fig. S5, E to H). In addition, somatic arm knockdown resulted in an increase in spectrosome-containing cells (Fig. 4J and fig. S5, C and D), reduced piwi expression in escort cells (Fig. 4F), and increased Bam expression in germ cells (Fig. 4G). Considering these results together, we propose that Btk29A acts on Arm, which in turn regulates piwi in the niche.

Arm functions in the canonical Wnt pathway (18). We therefore examined the ovaries of wg, Wnt2, Wnt4, and Wnt5 mutants; the germ cell overproduction was detected only in Wnt4 (fig. S5J). Somatic knockdown of Wnt4 aided by bab1-GAL4 resulted in a reduction in the expression of Piwi (Fig. 4H), accompanied by an accumulation of germ cells carrying spectrosomes (Fig. 4J and fig. S5, K to Q) with an increase in germline Bam expression (Fig. 4I). These findings support the hypothesis that Arm in the escort cells regulates germ cell proliferation under the control of Wnt4, which was likely derived from somatic cells other than cap cells and terminal filament cells, as hh-GAL4 selective for these cells was least effective to induce germ cell overproduction when used to drive Wnt4RNAi expression (fig. S5, K to Q).

To evaluate the ability of Arm to activate transcription, we used T cell factor (TCF) reporter assays with Cos7 cells transiently transfected with human Btk (hBtk) (fig. S3L). The wild-type hBtk alone was sufficient to induce phosphorylation at Y142 and Y654 of β-catenin (fig. S4E), whereas the kinase-dead hBtk (Btk-K430E) was not (fig. S4E). Tyrosine phosphorylation of β-catenin was completely blocked by two antagonists of hBtk (fig. S4E). Similarly, Btk29A type 2 phosphorylated Y142 and Y654 of mammalian β-catenin (fig. S4, B to D). Notably, the TCF reporter activity was six times as high when hBtk was transfected into Cos7 cells (fig. S3L) compared with the mock-transfected control, indicating that hBtk modulates the TCF-dependent transcriptional activation mechanism, in which Arm–β-catenin is involved as a coactivator (18).

We then examined the expression of an arm-dependent Ubx-lacZ reporter in the embryonic midgut. Btk29A knockdown abrogated the expression of this reporter (fig. S3, I to K), demonstrating that Btk29A supports Arm-dependent transcription in vivo.

We showed that Btk29A phosphorylates Arm–β-catenin on conserved tyrosine residues, one of which (Arm-Y150) is pivotal for the niche function to prevent GSC daughters from overproliferating (Fig. 4, D E, and J). Notably, most GSCs in Btk29A mutants do not express Bam (fig. S1R). This suggests that the presumptive Btk29A-Arm-Piwi pathway selectively regulates the proliferation of differentiating GSC daughters without interfering with GSC maintenance. Without Btk29A type 2, cystoblasts fail to exit the cell cycle, leading to the overproduction of germ cells, many of which are unable to complete differentiation and contribute to the genesis of an ovarian tumor. The model to explain how somatic Btk29A controls the switch from proliferation to differentiation in germ cells is shown in fig. S6.

β-Catenin exerts multiple functions through its promiscuous binding abilities in cell-to-cell interactions and transcription (18). This protein plays critical roles in stem cell biology, and β-catenin malfunction results in a variety of cancers (1). Our findings add a new dimension to the study of β-catenin by highlighting the pivotal role of the tyrosine phosphorylation of β-catenin in the control of transcription in the nucleus, in addition to the regulated control of the stability and motility of cell adhesion (17).

Supplementary Materials

Materials and Methods

Figs. S1 to S6

References (2025)

References and Notes

  1. Acknowledgments: We thank M. Asaoka, M. Bienz, J. J. Buggy, D. Drummond-Barbosa, P. Lasko, D. McKearin, K. Mochizuki, A. Nagafuchi, Y. Niki, M. Ote, A. Pélisson, K. Sato, H. Siomi, T. Tabata, Y. Tamori, R. Ueda, T. Uemura, H. White-Cooper, Bloomington Stock Center and Drosophila Genetic Resource Center for reagents, M. Suyama for secretarial assistance, and M. O. Gustafsson for technical help. This work was supported by Grant-in-Aids for Scientific Research (nos. 1802012, 23220007, and 24113502) from the Ministry of Education, Culture, Sports, Science, and Technology (Japan) to D.Y. B.F.N., and C.I.E.S. received support from the Swedish Cancer Society, the Swedish Research Council, and the Stockholm County Council (ALF).
View Abstract

Stay Connected to Science

Navigate This Article