Changes in rRNA Transcription Influence Proliferation and Cell Fate Within a Stem Cell Lineage

See allHide authors and affiliations

Science  17 Jan 2014:
Vol. 343, Issue 6168, pp. 298-301
DOI: 10.1126/science.1246384

Germline Pol I

RNA polymerase I (Pol I)–directed ribosomal RNA (rRNA) transcription has been extensively studied in mammalian cell lines and yeast. However, the functional significance of cell-specific regulation of Pol I transcription within developmental contexts in vivo remains unclear. Zhang et al. (p. 298) characterized a Drosophila Pol I regulatory complex and found that germline stem cells (GSCs) of the ovary exhibited increased levels of rRNA transcription relative to their immediate daughter cells. High levels of rRNA expression promoted GSC proliferation, with attenuation of Pol I activity showing effects during early germ cell differentiation.


Ribosome biogenesis drives cell growth and proliferation, but mechanisms that modulate this process within specific lineages remain poorly understood. Here, we identify a Drosophila RNA polymerase I (Pol I) regulatory complex composed of Under-developed (Udd), TAF1B, and a TAF1C-like factor. Disruption of udd or TAF1B results in reduced ovarian germline stem cell (GSC) proliferation. Female GSCs display high levels of ribosomal RNA (rRNA) transcription, and Udd becomes enriched in GSCs relative to their differentiating daughters. Increasing Pol I transcription delays differentiation, whereas reducing rRNA production induces both morphological changes that accompany multicellular cyst formation and specific decreased expression of the bone morphogenetic protein (BMP) pathway component Mad. These findings demonstrate that modulating rRNA synthesis fosters changes in the cell fate, growth, and proliferation of female Drosophila GSCs and their daughters.

Lineage-specific stem cell populations help to maintain tissues that experience high rates of cell turnover (1). Self-renewal and differentiation must be finely tuned to replace cells lost under normal physiological conditions and to rapidly compensate for acute cell loss. Although external cues from niches influence stem cell–based homeostasis (24), the intrinsic mechanisms that regulate differential growth and proliferation within stem cell lineages remain poorly understood.

We isolated the Drosophila recessive mutation under-developed1 (udd1) on the basis of its sterile phenotype. Staining for the germline markers Vasa and Hts (5, 6) revealed that udd1 mutants exhibit germ cell loss in ovaries and testes (figs. S1 and S2). Noncomplementation tests, reverse transcription–polymerase chain reaction (RT-PCR), and cDNA rescue experiments indicated that udd1 disrupts the expression of a divergent gene CG18316, referred to as udd hereafter (figs. S1 to S3). uddnull homozygotes exhibited embryonic lethality, which was rescued by expression of the udd open reading frame (fig. S2 and S3). Mosaic analysis revealed that udd1 and uddnull homozygous clones displayed egg-chamber degeneration similar to that of udd1/udd1 and udd1/uddnull mutants (fig. S1). Over time, udd1 and uddnull mutant GSCs became less proliferative and were eventually lost from the cap cell niche (Fig. 1).

Fig. 1 Disruption of udd results in reduced GSC proliferation and loss.

Negatively marked (A) control, (B) udd1 and (C) uddnull, clones (white dotted lines) dissected 21 days after clone induction stained for GFP (green), Hts (red), and DNA (blue). Arrowheads mark cap cells. (D) Percentage of germaria with GSC clones over time. Error bars indicate SD. (E) Percentage of GSC clones that produce a differentiating cyst over time. (F) Percentage of GSCs positive for the mitotic marker phospho-histone H3 (pH3). Scale bars: 10 μm.

Costaining with Modulo (Mod) (7) revealed that Udd protein exhibits ubiquitous expression and localizes to the nucleoli of nondividing cells (n > 100 cells) (Fig. 2A and fig. S3). Tandem purification and mass spectrometry (fig. S4), followed by coimmunoprecipitation (Fig. 2B and fig. S4), revealed that Udd associates with two proteins, CG6241 and CG10496. CG6241 shares homology with human TATA box–binding protein–associated factor RNA polymerase I subunit B (TAF1B) and yeast Rrn7 (8, 9) (fig. S5), whereas CG10496 resembles human TAF1C on the basis of sequence and secondary-structure analyses (fig. S6). CG6241 and CG10496 will hereafter be referred to as TAF1B and TAF1C-like, respectively. Human TAF1B and TAF1C are components of the selectivity factor 1 (SL1) complex, which promotes RNA polymerase I (Pol I) transcription (1012). Drosophila TAF1B and the TAF1C-like factor localized to nucleoli (Fig. 2, C and D, and fig. S4), and TAF1B was required for the localization and stability of Udd (fig. S7). Udd and TAF1B associated with the Pol I–specific subunit RpI135 (13) (fig. S4), and knockdown of TAF1B in the germ line resulted in phenotypes similar to that of udd1 (fig. S7).

Fig. 2 Characterization of a Drosophila SL1-like complex.

(A) w1118 germarium stained for endogenous Udd (green), Mod (red), and Vasa (blue). (B) Coimmunoprecipitation of FLAG-tagged TAF1B and TAF1C-like, and hemagglutinin (HA)–tagged Udd from transfected S2 cells. S2 cells transfected with constructs expressing (C) GFP-TAF1B and (D) TAF1C-like-GFP stained for GFP (green), Udd (red), and DNA (blue). Arrows mark transfected cells, and the arrowhead marks a nontransfected cell. (E) uddnull mosaics stained for Udd (green), BrUTP (red), and DNA (blue). Heterozygous nurse cells (arrowhead) exhibit Udd expression and BrUTP incorporation, whereas uddnull mutant cells (white-dotted line) show little BrUTP labeling. (F) Northern blot of total ovarian RNA isolated from the indicated genotypes probed with a fragment of internal transcribed spacer (ITS). Ethidium bromide–stained mature 28S and 18S rRNA. 5S rRNA was used as a loading control. (G) ChIP–quantitative PCR analysis of da>HA-udd ovaries reveals that Udd associates with specific sites within the rRNA promoter and external transcribed spacer (ETS), as indicated by arrows and bars. Control represents anti–HA immunoprecipitation from the da-gal4 background. Error bars represent SD. Scale bars: 10 μm (A and E); 5 μm (C and D).

To determine whether the Udd, TAF1B, and TAF1C-like complex promotes ribosomal RNA (rRNA) generation, we performed bromouridine-triphosphate (BrUTP) in situ run-on transcription assays to label nascent rRNA in uddnull clonal ovaries. BrUTP pulse-labeling revealed colocalization between nascent rRNA and Udd protein in control cells, but little BrUTP incorporation in homozygous uddnull mutant cells (Fig. 2E and figs. S8 and S9). RNA interference (RNAi) knockdown of TAF1B also reduced the synthesis of rRNA (fig. S7). Northern blot analysis (14) showed that udd mutants displayed a reduction in both pre-rRNA and processed rRNA intermediates (Fig. 2F and fig. S9). Chromatin immunoprecipitation (ChIP) experiments revealed that Udd associates with the rDNA promoter (Fig. 2G and fig. S9). Together these data indicate that the Udd, TAF1B, and TAF1C-like complex likely functions in a manner analogous to that of the human SL1 complex to promote Pol I transcription (fig. S7). As expected, disruption of Pol I transcription impeded ribosome production based on the nuclear accumulation of green fluorescent protein (GFP)–tagged RpS2 in udd1/uddnull mutant cells (fig. S9).

GSCs exhibited higher levels of rRNA synthesis and nucleolar Udd relative to their immediate progeny (Fig. 3A and fig. S10). These differences correlated with the expression of Bam, a key differentiation factor (1517) (Fig. 3B and fig. S10). Wicked, a component of the rRNA processing U3 small nucleolar ribonucleoprotein (snoRNP) complex, becomes enriched in cytoplasmic particles that asymmetrically segregate to presumptive GSCs during mitosis (18). To determine whether Udd also becomes asymmetrically enriched within GSCs, we performed immunofluorescence analysis of endogenous Udd and time-lapse microscopy using a rescuing GFP-tagged Udd genomic transgene (Fig. 3C, figs. S11 and S12, and movies S1 to S8). Live imaging showed discrete Udd-GFP localization during prophase. Udd-GFP dispersed during metaphase and anaphase, but a small amount of endogenous Udd remained associated with chromosomes through most of mitosis (fig. S11). At the end of telophase, GFP-tagged and endogenous Udd recoalesce within the nucleoli of newly formed GSCs more quickly and at higher levels relative to their sibling cells oriented away from the cap cells (Fig. 3C and fig. S11). By contrast, Udd appeared evenly distributed in multicellular cyst nucleoli immediately after mitosis (fig. S12).

Fig. 3 GSCs and undifferentiated cells maintain high levels of Udd.

(A) Control germarium stained for Udd (green), BrUTP (red), and DNA (blue). GSCs (inset) exhibit high levels of BrUTP labeling and Udd. (B) w1118 germarium stained for Bam (green), BrUTP (red), and DNA (blue). (C and C′) Still images from live-cell imaging (movies S1 and S2) showing GFP-tagged Udd (green) and monomeric red fluorescent protein (mRFP)–tagged histone H2Av (red) in a dividing GSC at the times indicated. Arrows point to the dividing GSC and resulting daughters. (D) No heat-shock (no HS) control hs-bam; bam∆86 mutant germarium and (E) heat-shocked (HS) hs-bam; bam∆86 mutant germarium 36 hours after bam induction stained for Udd (green), BrUTP (red) and DNA (blue). (D′ and E′) BrUTP labeling alone. Scale bars: 10 μm.

Udd and rRNA synthesis did not decrease in bam∆86 mutant germ cells (Fig. 3D and fig. S10), suggesting that persistently low levels of Pol I transcription during early cyst differentiation correlate with the developmental state of these cells and not with their position relative to the niche. To further explore this idea, we overexpressed an inducible bam transgene in a bam∆86 mutant background. Following bam expression, the germ cells differentiated into multicellular cysts, and both nucleolar Udd and nascent rRNA production levels decreased (Fig. 3E and fig. S10).

To examine the functional significance of reduced rRNA transcription in early differentiating cells, we crossed the udd1 mutation into a bam∆86 mutant background. Although bam∆86 mutant cells remained as single cells with round fusomes (Fig. 4A and fig. S13), udd1 bam∆86 double-mutant germaria (94.7%; n = 94) contained many four- and eight-cell cysts with branched fusomes and ring canals (Fig. 4B and fig. S13). Mature 16-cell cysts were not observed. RNAi knockdown of TAF1B in a bamRNAi background also resulted in multicellular cyst formation (fig. S13). Consistent with the idea that reduced translation promotes morphological changes that accompany early germline differentiation, knockdown of an rRNA processing factor, ribosomal proteins, and a translation initiation factor in a bam loss-of-function background also resulted in multicellular cyst formation (fig. S13). udd1 bam∆86 double-mutant germ cells maintain nucleolar fibrillarin and fail to down-regulate sex-lethal and up-regulate A2bp1 despite forming multicellular cysts (figs. S14 and S15).

Fig. 4 Modulating rRNA synthesis influences cyst development and BMP signaling in the germ line.

(A) bam∆86 and (B) udd1 bam∆86 double-mutant germaria stained for phosphotyrosine (pTyr) (green), Hts (red), and DNA (blue). Arrows point to cysts with branched fusomes. (C) nos-gal4 and (D) nos>Tif-IA stained for Hts (red) and DNA (blue). (E) Quantification of the number of single cells with round fusomes per germarium (germ) upon Tif-IA overexpression. (F) nos-gal4 and (G) nos>TIF-IA stained for Dad-LacZ. (H) Quantification of Dad-LacZ–positive cells upon Tif-IA overexpression. (I) Western blots of bam∆86 and udd1 bam∆86 ovarian extracts probed for Mad, Medea, and histone H2B proteins. (C to H) Two copies of nos-gal4 were present. (E and H) Error bars represent SEM. Scale bars: 20 μm. *P < 0.0001.

The udd bam double-mutant phenotype suggested that attenuation of Pol I activity promotes some of the early steps of germ cell differentiation. We speculated that increasing rRNA transcription in stem cell daughters exiting the niche might delay their ability to initiate cyst formation. Overexpression of TIF-IA, a conserved factor that bridges divergent Pol I regulatory factors with the Pol I transcriptional complex, results in greater rRNA transcription (19). Although we could not drive robust TIF-IA expression (fig. S16), low levels of TIF-IA overexpression resulted in a modest but significant increase in the number of single germ cells with round fusomes within germaria and the percentage of germaria containing over five single undifferentiated cells (Fig. 4, C to E). These cells continued to express Dad-LacZ, a hallmark of bone morphogenetic protein (BMP) signal transduction and GSC identity (20) (Fig. 4, F to H). We compared the levels of two downstream components of the BMP pathway, Mad and Medea (21), in bam∆86 and udd1 bam∆86 double mutants. Disruption of udd resulted in reduced levels of Mad but not Medea or histone H2B, indicating that modulation of rRNA transcription affects the expression of specific proteins that regulate cell-fate decisions within the GSC lineage (Fig. 4I and fig. S16). Down-regulation of Mad in response to reduced rRNA transcription likely acts in concert with other mechanisms that extinguish BMP signaling in GSC daughters displaced away from the stem cell niche (22, 23).

Besides TIF-IA and dMyc (19, 24), few regulators of Drosophila Pol I have been characterized. The identification of a Drosophila SL1-like complex provides insights into the mechanisms that regulate rRNA transcription in a developmental context (fig. S16D). Seminal work has shown that specific cellular structures asymmetrically segregate during stem cell divisions in Drosophila and mice (18, 2527). Results presented here indicate that rRNA transcriptional machinery also partitions unevenly during certain cell divisions. These data reveal that distinct levels of ribosome biogenesis, once considered a generally constitutive process, modulate the expression of specific proteins that direct cell fate decisions, growth, and proliferation within an in vivo stem cell lineage more rapidly or to a greater extent than others. Notably, the direction of asymmetric enrichment of ribosome biogenesis factors may be reversed in other lineages, especially in those stem cells destined to enter a quiescent state. These findings may have important implications for human ribosome-related diseases (28, 29).

Supplementary Materials

Materials and Methods

Figs. S1 to S16

References (3041)

Movies S1 to S8

References and Notes

  1. Acknowledgments: We thank L. Raftery, Z. Chen, A. Spradling, J. Pradel, N. Perrimon, the Bloomington Stock Center, and the Iowa Developmental Studies Hybridoma Bank for providing reagents. J. Huynh and P. R. Hiesinger provided imaging advice. N. Conrad, J. Jiang, and P. R. Hiesinger provided comments. Supported by NIH (R01GM086647 and R01GM045820) and E.E. and Greer Garson Fogelson Endowment (University of Texas Southwestern Medical Center).
View Abstract

Stay Connected to Science

Navigate This Article