Brat promotes stem cell differentiation via control of a bistable switch that restricts BMP signaling

Abstract

Drosophila ovarian germline stem cells (GSCs) are maintained by Dpp signaling and the Pumilio (Pum) and Nanos (Nos) translational repressors. Upon division, Dpp signaling is extinguished, and Nos is downregulated in one daughter cell, causing it to switch to a differentiating cystoblast (CB). However, downstream effectors of Pum-Nos remain unknown, and how CBs lose their responsiveness to Dpp is unclear. Here, we identify Brain Tumor (Brat) as a potent differentiation factor and target of Pum-Nos regulation. Brat is excluded from GSCs by Pum-Nos but functions with Pum in CBs to translationally repress distinct targets, including the Mad and dMyc mRNAs. Regulation of both targets simultaneously lowers cellular responsiveness to Dpp signaling, forcing the cell to become refractory to the self-renewal signal. Mathematical modeling elucidates bistability of cell fate in the Brat-mediated system, revealing how autoregulation of GSC number can arise from Brat coupling extracellular Dpp regulation to intracellular interpretation.

Figure 1

Brat Is Expressed in the Germline but Repressed by Pum-Nos in GSCs Germarium is indicated by the bracket in (A)–(D) and (G)–(H). Scale bars represent 10 μm (B–D) and 20 μm (G and H). (A) Structure of the Drosophila germarium and niche. TF, terminal filament; CC, cap cells; ESC, escort stem cell. (B) brat mRNA is expressed in the germarium, including GSCs (arrowheads). (C) Expression of Brat protein (red) is excluded from GSCs (arrowheads, spectrosome marked by α-Spectrin [green]). (D) Expression of Brat (red) is limited to CBs and differentiating cysts, marked by Bam-GFP expression (green). GSCs are labeled with arrowheads. (E) Representative western blot of GFP-brat3′ repression by Pum-Nos in S2 cells. RNAi was used to knock down endogenous Brat. (F) Quantification of data in (E) displayed as the average of three biological repeats. Error bars are +SEM; repression by Pum/Nos compared to no regulator is statistically significant (^∗p < 0.05). (G) Expression of Brat (red) in pum transheterozygous mutant ovaries, showing ectopic expression in GSCs (arrowheads). Ovaries were dissected immediately following eclosure to ensure the initial presence of germline cells. GSCs were identified by their position adjacent to niche cells and the fusome morphology and position. (H) As in (G), except that nos homozygous mutant ovaries were stained. (I) Table quantifying the presence of germline and number of GSCs ectopically expressing Brat protein in newly eclosed flies of different pum and nos mutant genotypes. Percentage of last column indicates proportion of Brat-positive GSCs out of total GSCs scored (n > 20). See also Figure S1.

Figure 2

Mutation of brat Increases Growth and GSC Number (A) brat¹¹ mutant germline, indicated by a loss of GFP (lower panel), shows an increase in cell size compared to wild-type (upper panel). (B) A brat¹¹ germarium with increased GSCs, indicated by extra fusomes (green, solid arrowheads) and lack of Bam expression (red). Bam-positive CBs are indicated (open arrowheads). (C) Quantitation of GSC number in brat¹¹ mutant (n = 20); error bars are +SEM (^∗p < 0.01). (D) Expression of Dad-LacZ (red) in wild-type and brat¹¹ mutant germaria, showing an expansion of LacZ-positive cells. (E) Quantitation displaying average number of Dad-lacZ-expressing germline cells (n = 6); error bars are +SEM (^∗p < 0.0)1. See also Figure S2.

Figure 3

Ectopic Brat Strongly Promotes Differentiation Images in (A)–(D) are compiled Z stacks. Scale bars represent 10 μm. (A) Wild-type germarium stained for Vasa (red) and α-Spectrin (green) to mark cytoplasm and fusome structures of germline cells, respectively. (B) Ectopic expression of Brat in the germline reduces number of germline cells, in addition to disruption of overall morphology of the ovary. Several germline-less ovarioles are present. (C) Ectopic Brat expression at higher levels through growth at 21°C enhances the loss of germline. (D) Ectopic expression at 25°C results in ovaries lacking germline cells. Analysis of cell death in Brat transgene expression using TUNEL assay to indicate apoptosis (red) (E)–(I) and viability staining to reveal other cell death, including necrosis (green) (E′)–(I′). Wild-type ovaries show little background staining for either cell death marker (E and E′), whereas ovaries from flies expressing the apoptosis-promoting gene grim under control of the heat shock promoter demonstrate widespread cell death following its induction (F and F′). Ovaries expressing the UAS-brat transgene initially show no cell death above background (G and G′), whereas after 7 days only apoptosing somatic cells are detected at the tip of ovarioles with maturing egg chambers, suggesting that they are remnants of germaria that have now emptied (H and H′, arrowheads). Following 21 days, most germaria are empty, and no egg chambers remain, leaving only apoptosing somatic cells (I and I′). No increase in germline cell death was detected at any stage. See also Figure S3.

Figure 4

Brat Functions Together with Pum in the Germline Images (A) and (B) are compiled Z stacks. Scale bars represent 10 μm. (A) Germaria of flies ectopically expressing Brat^R837D that cannot interact with d4EHP phenocopy wild-type ectopic expression, whereas those expressing Brat^G774D (B), which cannot interact with Pum, have a reduced phenotype, indicated by the presence of GSCs and a developing germline (Vasa, red). (C) Codetection of Brat and Pum shows that they overlap in CB cells (arrowhead). Separate channels show individual staining for clarity. (D) Codetection of Brat and Nos shows that they are expressed in reciprocal domains, suggesting that Pum-Brat activity is independent of Nos in differentiating cells. Separate channels are shown. See also Figure S3.

Figure 5

Brat Represses Mad and dMyc in the Ovary (A) Representative western blot showing Pum-Brat regulation of GFP-mad3′ in S2 cells. RNAi ensures that effects are independent of Nos. The intermediate level of repression by Pum is attributed to the contribution of endogenous Brat protein, and vice versa. Graph shows quantification of independent biological repeats (n = 4). Error bars are +SEM; repression by Pum-Brat compared to no regulator is statistically significant (^∗p < 0.05). (B and C) Western blot showing that GFP without 3′UTR (B) and epitope-tagged Tkv (C) are not regulated by Pum or Brat. (D) Pum-Brat regulation of GFP-dMyc3′ reporter, graph as in (A), repression by Pum-Brat compared to no regulator is statistically significant (^∗p < 0.05). (E) Expression of pMad (red) in wild-type (top) and brat mutant (bottom) germaria, showing expansion of activated Mad in germline cells mutant for Brat. (F) dMyc expression (red) in a brat mutant ovariole (bottom) begins earlier and is at a higher level compared to a wild-type ovariole (top). (G) Detailed view of dMyc expression (red, outlined) at the anterior of wild-type germaria, showing high level of expression in GSCs (top), and an expansion of this expression in the brat mutant (bottom). Scale bars represent 10 μm (E–G). See also Figure S4.

Figure 6

Brat-Mediated Repression of Mad and dMyc Leads to Bistability (A) Network diagram implemented in modeling studies, with components supporting self-renewal (yellow) or differentiation (blue). (B) Equilibrium distribution for pMad in wild-type, showing the two stable states (solid-black portions) and one unstable (dashed portion). Levels of pMad in Brat ectopic expression (lowest dashed line) and brat mutant (solid gray) illustrate the lack of bistability under these conditions. (C) Levels of pMad and other intracellular signal transducers and modifiers in the single-cell local model. (D) Comparison of network diagrams representing the “synthesis” model network (left) and the alternative “endocytosis” dMyc model (right). (E) Time course of the “synthesis” model. Colors indicate pMad levels (see color scale in next panel). From a wild-type germarium (t0), a clone expressing 4xdMyc is induced and allowed to stabilize (t1). Clone exhibits slightly higher levels of pMad, without affecting levels in adjacent cells. After a transient decrease in Dpp ligand, both the clone and wild-type GSC recover the self-renewal state, remaining as GSCs (t2). (F) Time course of the “endocytosis” model. The removal of local Dpp by the 4xdMyc cell lowers pMad signaling in an adjacent GSC (t1), causing this cell to be more susceptible to small perturbations in the level of Dpp, eventually leading to its differentiation (t2, asterisked cell). The remaining 4xdMyc GSC takes over the germarium (t3, arrow), demonstrating symmetrical division. (G) Model protein distributions for Brat, Bam, dMyc, and Mad in a wild-type germarium (see color scale). (H) pMad, dMyc, and Bam in the brat mutant (brat¹¹) and ectopic brat expression (brat oe) germaria, which mirror in vivo levels. For pMad wild-type, see (F, t0). (I) pMad expression during differentiation of GSCs predicted by the 3D spatiotemporal model. Immediately after division of a GSC (position 1), the pre-CB cell levels of pMad signaling drop (position 2), and signaling declines steadily and remains low in the two to four cell cysts (position 3). The newly formed CB follows trajectory 2, whereas the more posterior cyst follows trajectory 3 in the lower pMad state. Dedifferentiation of a cell, during repopulation of an emptied niche for example, is represented by the red line. (J) pMad expression during GSC differentiation in the 3D model of a brat mutant, representative of a system lacking bistability. pMad signaling is high in GSCs (position 1) and declines smoothly in the CB and two to four cell cysts (positions 2–4). See also Figure S5.

Figure 7

Brat's Role in the Ovarian Stem Cell System Schematic showing the function of Brat in the germline. See text for details.