Search tips
Search criteria 


Logo of hhmipaabout author manuscriptssubmit a manuscriptHHMI Howard Hughes Medical Institute; Author Manuscript; Accepted for publication in peer reviewed journal
Science. Author manuscript; available in PMC 2009 October 10.
Published in final edited form as:
PMCID: PMC2759887

Drosophila stem cells share a common requirement for the histone H2B ubiquitin protease Scrawny


Stem cells within diverse tissues share the need for a chromatin configuration that promotes self-renewal, yet few chromatin proteins are known to regulate multiple types of stem cells. We describe a Drosophila gene, scrawny (scny), encoding a ubiquitin protease, which is required in germline, epithelial and intestinal stem cells. Like its yeast relative UBP10, Scrawny deubiquitylates histone H2B and functions in gene silencing. Consistent with previous studies of this conserved pathway of chromatin regulation, scny mutant cells have elevated levels of ubiquitinylated H2B and trimethylated H3K4. Our findings suggest that inhibiting H2B ubiquitylation via scny represents a common mechanism within stem cells that is used to repress the premature expression of key differentiation genes, including Notch target genes.

Keywords: Stem cell, Ubiquitin specific protease, Histone H2B, gene silencing

Stem cells are maintained in an undifferentiated state by signals they receive within the niche and are subsequently guided toward particular fates upon niche exit (1). Within ES cells and during differentiation, cell state changes are controlled at the level of chromatin by alterations involving higher order nucleosome packaging and histone tail modifications (2). Polycomb group (PcG) and Trithorax group (trxG) genes influence key histone methylation events at the promoters of target genes, including H3K27 and H3K4 modifications associated with gene repression and activation, respectively, but few other genes with a specific role in stem cells are known.

Histone H2A and H2B mono-ubiquitylation play fundamental roles in chromatin regulation, and H2A ubiquitylation has been linked to PcG-mediated gene repression and stem cell maintenance. The mammalian Polycomb repressive complex 1 (PRC1) component RING1B is a H2A ubiquitin ligase that is required to block the elongation of poised RNA polymerase II on bivalent genes in ES cells (3). Mutations in the PRC1 component, BMI-1, which complexes with RING1B, causes multiple types of adult stem cells to be prematurely lost (4). The role of H2B ubiquitylation in stem cells is unclear, however. In yeast, ubiquitylation of Histone H2B by the RAD6 and BRE1 ligases controls H3K4 methylation (H3K4me3) (5,6), a process that requires the polymerase accessory factor PAF1 (7, 8). Conversely, H2B deubiquitylation by the ubiquitin-specific protease (USP) family member UBP10 is required for silencing telomeres, rDNA and other loci (9). The Drosophila homolog of BRE1, dBRE1, also is needed for H3K4 methylation, suggesting that this pathway is conserved (10). Furthermore, the Drosophila ubiquitin-specific protease USP7 is part of a complex that selectively deubiquitylates H2B and genetically interacts with PcG mutations (11). Mutations in another USP family member, Nonstop, increase H2B ubiquitylation and cause axon targeting defects in the eye (12).

In order to gain further insight into the role of H2B ubiquitylation in stem cells, we characterized a novel Drosophila gene, scrawny (scny) (CG5505), whose encoded USP family protein shares homology with human USP36 and among yeast USPs closely matches UBP10 within the core protease domain (13, Fig. 1A). Strains bearing scny insertions (Fig. 1B), except for a viable GFP protein trap (CA06690), were female sterile or lethal, and proved to be allelic (13, Table S1). Transposon excision or expression of a scny-RB cDNA reverts the phenotype of tested alleles. An anti-SCNY antibody raised against a domain common to all SCNY isoforms recognizes wild type and SCNY-GFP on a Western blot (Fig. 1C). SCNY protein levels in homozygous third instar larvae are greatly reduced in lethal mutants (Fig. 1C), and SCNY expression is also lower in stem cell-enriched ovarian tissue from adults homozygous for the sterile d06513 allele (Fig. 1D). Consistent with a role in gene silencing, several scny mutations act as dominant suppressors of position effect variegation (Fig. 1E, Fig. S1).

Figure 1
scny encodes a ubiquitin protease family member and functions in gene silencing. (A) Schematic of SCNY protein shows similarity to human USP36 and yeast UBP10 in the conserved protease (C19E) domain (red boxes). Blue bar indicates region used for antibody ...

Further studies strongly suggested that SCNY functions in vivo as an H2B-ubiquitin protease. Recombinant full-length SCNY protein, but not a version bearing a point mutation in the protease domain, efficiently deubiquitylates histone H2B in vitro (Fig. 2A). scnyf01742 homozygous tissue contains levels of Ub-H2B that are elevated at least twofold compared to wild type (Fig 2B, Fig. S2). As expected if Ub-H2B is required for H3K4 methylation, clones of homozygous scnye00340 mutant cells stain more strongly for H3K4me3 than heterozygous cells (Fig. 2C, C′). Consistent with a direct rather than an indirect action on Ub-H2B levels, anti-SCNY antibodies but not pre-immune serum co-immunoprecipitated H2B from Drosophila embryonic nuclear extracts (Fig. 2D). Moreover, epitope-tagged SCNY co-immunoprecipitated Drosophila PAF1, but not Cyclin T (or several other tested chromatin proteins) when co-expressed in S2 tissue culture cells (Fig. 2E). Together, these data support the view that SCNY participates in a conserved pathway of chromatin regulation linking H2B ubiquitylation with H3K4me3 methylation. Because the effects of scny mutation on Ub-H2B and H3K4me3 are opposite to those of dBre1 mutation (10), SCNY likely opposes dBRE1 action on H2B, just as UBP10 opposes BRE1 action on H2B in yeast.

Figure 2
SCNY deubiquitylates H2B in vitro, affects chromatin modification in vivo, and interacts with H2B and PAF1. (A) Wild-type SCNY protein (SCNY) but not SCNY protein bearing a protease domain mutant (SCNYut) deubiquitylates Drosophila embryonic histone H2B ...

Drosophila male and female gonads contain well characterized germline stem cells (GSCs) that allow the effects of genes on stem cell maintenance to be quantitatively analyzed (14). High levels of scny expression were observed in female and male GSCs using SCNY-GFP (Fig. 3A, B) and identical staining was observed using anti-SCNY immunofluoresence (Fig. S3). SCNY protein resides in cell nuclei and is enriched in nucleoli (Fig. S3). In sterile or semi-fertile scny mutant adults, the numbers of germline stem cells surrounding the testis hub (Fig. 3C, Fig. S4) and within germaria (Fig. S5) were clearly reduced. The half-lives of female GSCs bearing clones of three different scny alleles were all sharply reduced (Fig. 3D). Later follicular development was also abnormal suggesting that scny continues to function after the stem cell stage. However, previous studies indicate that accelerated GSC loss is a specific phenotype, and hence that scny has a preferential requirement in GSCs (1, 15).

Figure 3
SCNY maintains germline stem cells. Wide expression of scny-GFP (green) during early oogenesis (A) and spermatogenesis (B), including GSCs (arrows). SCNY protein localizes to the nucleus and is highly enriched in the nucleolus. HTS (red). Hub (dashed ...

A known mechanism of increased GSC loss is the premature activation of differentiation genes. Staining germaria with an antibody specific for multiple sites of histone H3 acetylation (H3-Ac) suggested that scny mutation affects the global chromatin organization of GSCs (Fig. 3E, F). Wild type GSCs (arrow) contain lower levels of H3-Ac than slightly older germ cells within cysts (Fig. 3E, arrowhead). Presumptive GSCs located in the GSC niche in scny mutants frequently stained more strongly (Fig. 3F), suggesting that they have begun to upregulate general transcription. Some scny GSC-like cells also expressed bag-of-marbles (bam) (Fig. S6), a key cystoblast differentiation gene (1), and GSC-like cells in scnyd06513; bamΔ86 mutant females persist in the germarium (Fig. S7). However, we could not completely rule out that the observed increases in H3-Ac levels and bam expression were a result rather than a cause of the premature differentiation and loss of scny GSCs.

To determine if scny is also required in a very different type of stem cell, the epithelial FSC (16), we quantitatively analyzed the persistence of individual scny mutant FSCs. The half-life of FSCs mutant for scnyl(3)02331 was reduced more than 10-fold, while the scnyf01742 mutation also caused a sharp decline (Fig. 4A). However, mutant follicle cells continued to develop normally at later stages (Fig. S8). Thus, scny is preferentially required to maintain FSCs as well as GSCs.

Figure 4
SCNY maintains FSCs and ISCs. (A) Reduced halflife (t1/2) of FSCs mutant for scny. (B) Expression of SCNY-GFP (green) in nuclei and nucleoli of ISCs (arrow, contains cytoplasmic Delta), and developing enterocytes (lacks Delta), but is low in enteroendocrine ...

The largest population of Drosophila stem cells are the hundreds of multipotent intestinal stem cells (ISCs) that maintain the adult posterior midgut (17, 18). ISCs signal to their daughters via Delta-Notch signaling to specify enterocyte vs. enteroendocrine cell fate, but the pathway must remain inactive in the ISCs themselves to avoid differentiation (19). Most ISCs (those about to produce enterocytes) express high levels of the Notch ligand Delta, allowing them to be specifically distinguished from other diploid gut cells (19). We found that SCNY-GFP is expressed in ISCs (Fig. 4B, arrow) suggesting that SCNY plays a role in these stem cells as well. While 7-day old normal adult midguts contain a high density of ISCs, as revealed by Delta staining (Fig. 4C), we found that corresponding tissue from 7-day-old scnyf01742 (Fig. S9) or scnyf01742/scnyl(3)02331 (Fig. 4D) escaper adults possess very few Delta-positive cells. ISCs are present in near normal numbers at eclosion, but are rapidly lost in the mutant adults, indicating that scny is required for ISC maintenance (Fig. 4F).

We suspected that inappropriate Notch pathway activation was responsible for the premature ISC loss in scny mutants. dBre1 mutations strongly reduce Notch signaling, suggesting that Notch target genes are particularly dependent on H2B mono-ubiquitylation and H3K4 methylation (10). Consequently, scny mutations, which have the opposite effects on Ub-H2B and H3K4me3 levels, might upregulate Notch target genes, stimulating ISCs to differentiate prematurely. We tested this idea by supplementing the food of newly eclosed scnyf01742/scnyl(3)02331 adults with 8 mM DAPT, a gamma-secretase inhibitor that blocks Notch signaling and phenocopies Notch mutation when fed to wild type animals (17). scnyf01742/scnyl(3)02331 DAPT-treated adults remained healthy and the guts of 7-day old animals still contained many ISCs (Fig 4E), although not as many as wild type (Fig, 4F). Tumors like those produced in wild type animals fed DAPT (17, 18) were not observed. Thus, in these animals endogenous stem cell loss can be slowed by drug treatment.

Our experiments provide strong evidence that a pathway involving the ubiquitin protease Scrawny and the ubiquitin ligase dBRE1 controls the levels of Ub-H2B, and H3K4me3 at multiple target sites in the Drosophila genome. Although, other ubiquitin proteases also act on Ub-H2B in Drosophila (11, 12), the direct interaction between SCNY and H2B, and the strong effects of scny mutations argue that it plays an essential, direct role in silencing genomic regions critical for cellular differentiation, including Notch target genes. SCNY interacts with the RNA polymerase accessory factor complex component, PAF1. Upregulation of H2B ubiquitinylation and H3 methylation in yeast is mediated by the PAF1 complex and is associated with elongating RNA Pol II (7, 8, 20). Drosophila PAF1 is required for normal levels of H3K4me3 at the hsp70 gene (21), and another PAF1 complex member, RTF1, is needed for H3K4 methylation and Notch target gene expression (22). Indeed, the pathway connecting Ub-H2B, H3K4me3 and gene silencing appears to be conserved in organisms as distant as Arabidopsis (23). A human protein closely related to SCNY, USP36, is overexpressed in ovarian cancer cells (24), and our results suggest it may act as an oncogene by suppressing differentiation.

Above all, our experiments indicate that SCNY-mediated H2B deubiquitylation is required to maintain multiple Drosophila stem cells, including progenitors of germline, epithelial and endodermal lineages. In ES cells and presumably in adult stem cells, many differentiation genes contain promoter-bound, arrested RNA Pol II and are associated with Polycomb group proteins (2, 3). We envision that in the niche environment SCNY activity overrides that of dBRE1, keeping levels of Ub-H2B (and hence H3K4me3) low at key differentiation genes (Fig. S10). Upon exit from the niche, the balance of signals shifts to favor H2B ubiquitylation, H3K4 trimethylation, and target gene activation. Thus, the control of H2B ubiquitylation, like H2A ubiquitylation, plays a fundamental interactive role in maintaining the chromatin environment of the stem cell state.

Supplementary Material



Supporting Online Material

Materials and Methods

Figs. S1 to S10

Table S1



1. Morrison S, Spradling AC. Cell. 2008;132:598. [PubMed]
2. Spivakov M, Fisher AG. Nat Rev Genet. 2007 Apr;8:263. [PubMed]
3. Stock JK, et al. Nat Cell Biol. 2007;9:1428. [PubMed]
4. Valk-Lingbeek ME, Bruggeman SW, van Lohuizen M. Cell. 2004 Aug 20;118:409. [PubMed]
5. Sun ZW, Allis CD. Nature. 2002 Jul 4;418:104. [PubMed]
6. Dover J, et al. J Biol Chem. 2002 Aug 9;277:28368. [PubMed]
7. Wood A, Schneider J, Dover J, Johnston M, Shilatifard A. J Biol Chem. 2003 Sept 12;278:34739. [PubMed]
8. Ng HH, Dole S, Struhl K. J Biol Chem. 2003 Sep 12;278:34739. [PubMed]
9. Gardner RG, Nelson ZW, Gottschling DE. Mol Cell Biol. 2005 Jul;25:6123. [PMC free article] [PubMed]
10. Bray S, Musisi H, Bienz M. Dev Cell. 2005 Feb;8:279. [PubMed]
11. van der Knaap JA, et al. Mol Cell. 2005 Mar 4;17:695. [PubMed]
12. Weake VM, et al. EMBO J. 2008 Jan 10;27:394. [PubMed]
14. Xie T, Spradling AC. Cell. 1998 July 24;94:251. [PubMed]
15. Lin H, Spradling AC. Development. 1997 June;124:2463. [PubMed]
16. Nystul TG, Spradling AC. Cell Stem Cell. 2007;1:277. [PubMed]
17. Ohlstein B, Spradling AC. Nature. 2006 Jan 26;439:470. [PubMed]
18. Micchelli C, Perrimon N. Nature. 2006 Jan 26;439:475. [PubMed]
19. Ohlstein B, Spradling AC. Science. 2007 Feb 16;315:988. [PubMed]
20. Xiao T, et al. Mol Cell Biol. 2005 Jan;25:637. [PMC free article] [PubMed]
21. Adelman K, et al. Mol Cell Biol. 2006 Jan;26:250. [PMC free article] [PubMed]
22. Tenney K, et al. Proc Natl Acad Sci USA. 2006 August 8;103:11970. [PubMed]
23. Wang H, et al. Nature. 2004 Oct 14;431:873. [PubMed]
24. Li J, et al. Int J Med Sci. 2008 June 6;5:133. [PMC free article] [PubMed]
25. This work was supported by the Howard Hughes Medical Institute. M.B. was supported in part by American Cancer Society Followship #PF-04-022-01-CSM. We would like to thank Dennis McKearin, Moshe Oren and the Bloomington Stock Center for reagents. We would like to thank Ben Ohlstein for helpful advice and assisting with the initial characterization of the scny midgut phenotype. We would like to thank Todd Nystul for critical reading of the manuscript.