Loss of Drosophila Ataxin-7, a SAGA subunit, reduces H2B ubiquitination and leads to neural and retinal degeneration

Abstract

The Spt-Ada-Gcn5-acetyltransferase (SAGA) chromatin-modifying complex possesses acetyltransferase and deubiquitinase activities. Within this modular complex, Ataxin-7 anchors the deubiquitinase activity to the larger complex. Here we identified and characterized Drosophila Ataxin-7 and found that reduction of Ataxin-7 protein results in loss of components from the SAGA complex. In contrast to yeast, where loss of Ataxin-7 inactivates the deubiquitinase and results in increased H2B ubiquitination, loss of Ataxin-7 results in decreased H2B ubiquitination and H3K9 acetylation without affecting other histone marks. Interestingly, the effect on ubiquitination was conserved in human cells, suggesting a novel mechanism regulating histone deubiquitination in higher organisms. Consistent with this mechanism in vivo, we found that a recombinant deubiquitinase module is active in the absence of Ataxin-7 in vitro. When we examined the consequences of reduced Ataxin-7 in vivo, we found that flies exhibited pronounced neural and retinal degeneration, impaired movement, and early lethality.

Keywords: H2B ubiquitination; SAGA complex; spinocerebellar ataxia.

Figure 1.

The putative Drosophila homolog of Ataxin-7 is CG9866. (A) Analysis of Drosophila SAGA complexes by MudPIT mass spectrometry consistently identified the protein product of the uncharacterized gene CG9866 (Ataxin-7 hereafter). A comparison of complex normalized spectral abundance factor (cNSAF) values for the SAGA subunits found in purifications performed with the indicated bait proteins is shown (Weake et al. 2009). (B) Human and Drosophila Ataxin-7 share a high degree of primary sequence homology in the Ataxin-7 signature blocks defined by Helmlinger et al. (2004). Amino acid sequences from human and Drosophila Ataxin-7 were aligned using ClustalW and shaded using Box Shade. Identical residues are shown with a black background and white text; similar residues are shown with a light-gray background and black text. The complete alignment can be found in Supplemental Figure S1.

Figure 2.

Ataxin-7 is a member of the SAGA chromatin-modifying complex. (A) Ataxin-7-containing complexes were affinity-purified from S2 cells stably transfected with an expression vector for Ataxin-7-HF and selected to express low levels of recombinant protein. Silver staining shows the protein profile of the purified complex, including the expected position of bait proteins. Complexes purified through another SAGA subunit (HA-Flag-WDA) are included for comparison. (B) Immunoblotting of purified Ataxin-7 complexes confirmed the presence of SAGA complex members in the purification. (C) MudPIT proteomic analysis verified that the entire SAGA complex was present in the purification. The predicted molecular mass of each protein is indicated. (D) Purified Ataxin-7-containing complexes were fractionated by size on a Superose 6 gel filtration column. Eluted fractions were analyzed by immunoblotting with the indicated antibodies to detect Ataxin-7 (αHA) and SAGA-specific complex member Ada2B (αAda2B). The majority of Ataxin-7 stably interacted with SAGA, coeluting with Ada2B.

Figure 3.

A P-element promoter insertion reduces Ataxin-7 levels in vivo. (A) The organization of the Ataxin-7 gene and the location of a P-element insertion in the 5′ UTR are shown. (B) Larvae homozygous for the P-element insertion have drastically reduced levels of Ataxin-7 protein. Whole-cell extracts were prepared from wandering third instar larvae and analyzed by immunoblotting with an Ataxin-7-specific antibody. (Bottom panel) Equal loading of protein extracts was verified by probing for β-tubulin. (Right panel) The reduction in Ataxin-7 expression was quantified, and the standard error is indicated. (C) Analysis of polytene chromosomes showed colocalization between Ataxin-7 and Pol II (wild type [WT]). A drastic reduction in detectable Ataxin-7 was seen on chromatin from (P)Ataxin-7 mutant larvae.

Figure 4.

The enzymatic modules of SAGA are separated upon loss of Ataxin-7. (A) SAGA complex composition is altered by loss of Ataxin-7. Nuclear extracts were prepared from wandering third instar Ataxin-7 mutant (MT) or wild-type OregonR (WT) larvae. (Top four panels) Similar amounts of extract, as quantified by Bradford assay, were fractionated by size on a Superose 6 gel filtration column as above, and aliquots of the eluted fractions were analyzed by immunoblotting with an Ada2B-specific antibody to detect SAGA and confirm equal loading of the column. (Bottom four panels) Immunoblotting with an Ataxin-7-specific antibody showed that endogenous Ataxin-7 coeluted with high-molecular-weight complexes. (B) The association between Non-stop and Gcn5 is reduced in the absence of Ataxin-7. Endogenous Non-stop-containing complexes were immunoprecipitated from whole-cell extracts prepared from wild-type (WT) or Ataxin-7 mutant third instar larvae using a Non-stop-specific antibody or preimmune serum (control). (Left panels) Equivalent levels of Non-stop and Gcn5 in both extracts and decreased levels of Ataxin-7 in the mutant extract were verified by immunoblotting. (Top right panel) The presence of Gcn5 in the immunoprecipitate was probed using a Gcn5-specific antibody. (Bottom right panel) Equal capture of Non-stop was verified by immunoblotting for Non-stop. (C) Endogenous SAGA-independent Non-stop-containing complexes are enzymatically active. The Non-stop immunoprecipitates from B were assayed using the ubiquitin-AMC assay. Arbitrary units of fluorescence, subtracted for background in the control immunoprecipitation, are shown, and the standard error is indicated.

Figure 5.

Loss of Ataxin-7 results in a Non-stop-dependent reduction of H2B ubiquitination. (A) Non-stop maintains chromatin localization in the absence of Ataxin-7. Polytene chromosome squashes prepared from Ataxin-7 mutant third instar larvae were immunostained with a Non-stop-specific antibody. (B) H2B ubiquitination and H3K9 acetylation are reduced in the absence of Ataxin-7. Bulk histones acid-extracted from OregonR, Ataxin-7 mutant, or hop-out (Rescue) third instar larvae were analyzed by immunoblotting for the indicated histone modifications. H4 served as a loading control, and representative blots are shown. This analysis indicated a significant decrease in H2B ubiquitination and H3K9 acetylation in Ataxin-7 mutants that was rescued by precise excision of the P-element. The levels of H2Bub quantified from at least three biological repeats are shown in the right panel. The standard error is indicated. (C) The lethality of Ataxin-7 mutants was relieved by reducing Non-stop. Flies of the indicated genotypes were mated, and progeny were grown under standard culture conditions. The genotypes of the surviving adult progeny are indicated. Ataxin-7 homozygous mutants did not survive to adulthood (n = 1255). However, reducing one copy of Non-stop partially relieved this lethality, with 12% of the adult progeny being homozygous for the Ataxin-7 mutation and heterozygous for the (P)Non-stop mutation. Non-stop homozygous mutants do not survive to adulthood, irrespective of the presence or absence of mutations in Ataxin-7.

Figure 6.

The effect of Ataxin-7 knockdown on H2B ubiquitination is conserved in higher eukaryotes. (A) Ataxin-7 expression was knocked down in HeLa cells by transfection with two different siRNAs. Forty-eight hours post-transfection, cells were harvested, and whole-cell extracts were prepared. Analysis of these extracts by immunoblotting showed decreased levels of H2Bub upon knockdown of Ataxin-7. Quantification of three independent experiments is shown at the right, and the standard error is indicated. (B) The effect of Ataxin-7 knockdown on H2Bub levels is rescued by hSgf11 knockdown. The USP22 deubiquitinase was inactivated by knocking down hSgf11 in HeLa cells, and levels of H2Bub increased as expected. Knockdown of Ataxin-7 and hSgf11 together also resulted in an increase in H2Bub, indicating that active USP22 is needed for the reduction of H2Bub that occurs upon Ataxin-7 knockdown.

Figure 7.

The SAGA deubiquitinase module is enzymatically active in the absence of Ataxin-7. (A) SAGA complex purified through Ataxin-7 was assayed for deubiquitinase activity using the reporter substrate ubiquitin-AMC. Increasing amounts of Ataxin-7-containing complex were incubated with an excess of ubiquitin-AMC, and release of AMC from ubiquitin was measured by indirect fluorescence. Time- and dose-dependent deubiquitination was observed. (B) Schematic of the experimental flow for analysis of deubiquitinase-containing complexes: SAGA complex purified through Ataxin-7 was separated by size on a gel filtration column, and fractions were assayed for enzymatic (deubiquitinase) activity. (C) Purified complex was fractionated by size, and fractions were assayed for enzymatic activity. Peak activities detected by ubiquitin-AMC activity assay are outlined by dotted lines, indicating that Ataxin-7 associates with intact and enzymatically active SAGA complex. (D, left panel, inset) A recombinant deubiquitinase module was isolated by coinfecting SF9 cells with baculovirus expression vectors for Flag-HIS-Sgf11, HA-E(y)2, HA-Non-stop, and HA-CG9866 (1–200), as indicated. (Right panel) The recombinant deubiquitinase module was then purified from whole-cell lysates through the Flag epitope present on Flag-HIS-Sgf11, and the components of the reaction were verified by immunoblotting to ensure complex integrity. (Left panel) These complexes were assayed for enzymatic activity using the ubiquitin-AMC assay.

Figure 8.

Loss of Ataxin-7 results in neural and retinal degeneration, reduced locomotion, and shortened life span. (A) The majority of (P)Ataxin-7 mutants do not survive to adulthood, but a small number of escapers survive under careful culture conditions. We analyzed the heads of these mutants by histological sectioning followed by hematoxylin and eosin (H&E) staining and analysis by bright-field microscopy. Widespread lesions were apparent throughout the mutant head. Similarly, knockdown of Ataxin-7 by actin-Gal4-directed ubiquitous expression of RNAi resulted in lesions throughout the head, similar to those seen in mutants. The mutant was rescued by excision of the P-element, restoring neural and retinal integrity. (B) Targeted knockdown of Ataxin-7 within the retina (R) and lamina (L) by GMR-Gal4-driven Ataxin-7-specific RNAi resulted in a progressive age-dependent rough-eye phenotype. Eyes were visualized by electron microscopy. (C) Two weeks post-eclosion, fly heads from GMR-GAL4-driven RNAi knockdown were analyzed as in A. Morphological analysis showed severe deterioration of the targeted tissues but normal physiology in untargeted areas. (D) The surviving homozygous (P)Ataxin-7 flies displayed locomotor defects in a negative geotaxis assay. In contrast to the rapid geotaxis behavior of wild-type flies, these mutant flies were unable to scale the side of a vial within 20 sec, and mobilization of the P-element restored climbing ability. (E) Survival curves show that mutant flies have a shortened life span compared with wild-type counterparts, and normal life span was restored by mobilization of the P-element.

Figure 9.

Model. Evidence from experiments performed in S. cerevisiae and here in Drosophila suggest that SAGA complex members associate in a modular fashion, with the yeast Ataxin-7 homolog Sgf73 linking modules containing the dual enzymatic activities acetyltransferase (Gcn5/Pcaf) and deubiquitinase (Ubp8/Non-stop/Usp22). In higher eukaryotes, the deubiquitinase module may function in the absence of Ataxin-7.