Functional organization of the yeast SAGA complex: distinct components involved in structural integrity, nucleosome acetylation, and TATA-binding protein interaction

Abstract

SAGA, a recently described protein complex in Saccharomyces cerevisiae, is important for transcription in vivo and possesses histone acetylation function. Here we report both biochemical and genetic analyses of members of three classes of transcription regulatory factors contained within the SAGA complex. We demonstrate a correlation between the phenotypic severity of SAGA mutants and SAGA structural integrity. Specifically, null mutations in the Gcn5/Ada2/Ada3 or Spt3/Spt8 classes cause moderate phenotypes and subtle structural alterations, while mutations in a third subgroup, Spt7/Spt20, as well as Ada1, disrupt the complex and cause severe phenotypes. Interestingly, double mutants (gcn5Delta spt3Delta and gcn5Delta spt8Delta) causing loss of a member of each of the moderate classes have severe phenotypes, similar to spt7Delta, spt20Delta, or ada1Delta mutants. In addition, we have investigated biochemical functions suggested by the moderate phenotypic classes and find that first, normal nucleosomal acetylation by SAGA requires a specific domain of Gcn5, termed the bromodomain. Deletion of this domain also causes specific transcriptional defects at the HIS3 promoter in vivo. Second, SAGA interacts with TBP, the TATA-binding protein, and this interaction requires Spt8 in vitro. Overall, our data demonstrate that SAGA harbors multiple, distinct transcription-related functions, including direct TBP interaction and nucleosomal histone acetylation. Loss of either of these causes slight impairment in vivo, but loss of both is highly detrimental to growth and transcription.

FIG. 1

Western blot characterization of Superose 6-purified spt3Δ and spt8Δ SAGA complexes. (A) SAGA complexes derived from wild-type (FY631), spt3Δ (FY294), and spt8Δ (FY462) strains were pooled, concentrated, and run on a Superose 6 column for separation based on molecular weight. Western blots of these fractions were performed to compare the sizes of the complexes. Blots were visualized with dilutions of antisera raised against the Ada2 or Spt3 proteins as indicated. (B) Western blots of wild-type (fraction 20), spt3Δ (fraction 21), and spt8Δ (fraction 21) SAGA, visualized with antisera specific to Spt8, Taf_II90, Taf_II68, or Taf_II60.

FIG. 2

Presence of Ada1 in the SAGA complex. HAT complexes were derived from wild-type (PSY316) and ada1Δ (SB11) yeast strains through the Mono Q column step of the previously described purification procedure. (A) Fluorographs of nucleosome acetylation assays of the even-numbered column fractions. The arrows denote the relative positions of the four histone proteins as separated on SDS–18% polyacrylamide gels; visualized species contain acetyl groups labeled with ³H. Wild-type fractions displaying the H3/H2B-acetylating activities of the Ada and SAGA complexes are indicated at the top. (B) Western blots of the fractions, visualized with dilutions of antisera raised against Ada1, Ada2, or Gcn5 protein.

FIG. 3

Phenotypic comparison of ada1Δ and other SAGA deletion mutants. Strains FY630, FY1106, FY1559, FY297, FY463, FY1599, and FY1553 were grown overnight in liquid YPD medium. For each strain, approximately 5 × 10³ cells were spotted on the indicated plates and grown for 2 days at 30°C. All strains contain the his4-917δ insertion mutation. Otherwise, wild-type strains containing this allele are His⁻, while strains containing a mutation able to suppress this allele are His⁺ (Spt⁻ phenotype).

FIG. 4

Comparison of gcn5Δ spt3Δ and gcn5Δ spt8Δ double mutants with an spt20Δ mutant. (A) Strains FY2, FY293, FY1299, FY1286, FY1440, FY1719, and FY1098 were grown overnight in liquid YPD medium. For each strain, approximately 2 × 10³ cells were spotted on the indicated plates. Growth is shown after 2 days of incubation at 30°C, except in the case of the caffeine plates, which were photographed after 3 days. (B) Double-mutant SAGA complexes are intact. Mono Q fractions from gcn5Δ spt3Δ (FY1441) and gcn5Δ spt8Δ (FY1720) HAT complex preparations were analyzed by Western blotting. SAGA subunits were visualized with dilutions of antisera raised against Ada2 or Spt20 protein as indicated.

FIG. 5

Effect of Gcn5 bromodomain deletion on nucleosomal HAT activity. Cells containing wild-type (w.t.; residues 1 to 439) or ΔBrD (residues 1 to 350 or 1 to 280) GCN5 were used to prepare Mono Q fractions by the standard purification method for the HAT complexes. (A) HAT assays with nucleosomal or free histone substrates were performed with even-numbered fractions surrounding the Ada and SAGA peaks; fluorographs of the resulting gels are presented. The arrows denote the relative positions of the four histone proteins as separated on SDS–18% polyacrylamide gels; visualized species contain acetyl groups labeled with ³H. The bar graph at right presents the ratio of nucleosomal to free histone H3/H2B acetylation for SAGA fraction 40, normalized to the wild-type ratio. (B) Samples of fractions 20 (Ada peak), 30 (negative control; no HAT complex), and 40 (SAGA peak) were also analyzed by Western blotting. Five microliters of each was run on SDS-polyacrylamide gels and electroblotted to nitrocellulose, which underwent immunodetection with dilutions of anti-Ada2, anti-Gcn5, or anti-Taf_II antibodies. Markers on the Gcn5 panels indicate the Gcn5 species visualized on equivalent portions of the blots, and their positions of migration agreed with their predicted sizes (51, 40, and 32 kDa for the 1-439, 1-350, and 1-280 constructs, respectively). The peak SAGA fraction from a preparation of gcn5Δ (FY1370) cells was also tested for Taf_IIs (right).

FIG. 6

HIS3 transcriptional defects observed in GCN5 null, HAT domain substitution, and ΔBrD mutants. (A) Diagram of HIS3 promoter function. Weak and strong TBP-binding sequences (TATA boxes) direct transcription from the +1 and +13 start sites, respectively. The activator Gcn4 binds to multiple upstream sites (UAS). (B) Start site usage for activated HIS3 transcription in wild-type (w.t.), gcn5Δ, and GCN5 HAT domain substitution mutant strains. Activating conditions were achieved with 3-aminotriazole. PhosphorImager quantitation of the S1 assays is presented as the ratio of +13 to +1. Shown at the bottom are in vitro HAT activities (as a percentage of wild-type activity) of native SAGA complexes purified from the indicated strains (27). (C) Start site preference for wild-type, gcn5Δ, and ΔBrD strains under noninducing conditions (no 3-aminotriazole). (D) Activated HIS3 transcription in wild-type, gcn5Δ, and ΔBrD strains. PMA1 RNA and tRNA were included as internal controls for sample loading.

FIG. 7

In vitro binding of wild-type, spt3Δ, and spt8Δ SAGA complexes to TBP. Glutathione-Sepharose beads containing bacterially expressed GST or GST-TBP were incubated alone or with similar amounts of wild-type, spt3Δ, or spt8Δ SAGA complexes at 4°C under conditions of approximately 60 mM NaCl. After washing, Western blotting was performed with these beads to analyze the proteins that had bound to GST (G lanes) or GST-TBP (T lanes). The input (I) lanes contained half of the amount of SAGA used for the binding experiments. The position of the antibody-visualized Ada2, Spt20, and Spt3 bands on the Western blots are indicated by arrows. An unidentified species in the GST-TBP preparation (T lanes) had a mobility similar to but slightly slower than that of Spt3 and cross-reacted with anti-Spt3 antibodies; the band seen in the spt8Δ T lane is a combination of this species and a small amount of Spt3.

FIG. 8

Model for SAGA structure and function in transcription. Depicted is a hypothetical gene with an upstream activation sequence (UAS) and TATA box; the DNA is wound around nucleosomes (cylinders). The SAGA complex, composed of adaptor (white) and Spt (black) functional regions and held together by several structurally important proteins (grey), is proposed to interact with the an activator (such as Gcn4) at the UAS through Ada2. This allows the HAT activity of Gcn5, in cooperation with its bromodomain (BrD), to acetylate (Ac = acetyl group) the amino-terminal tails of nucleosomal histones, providing more effective TATA binding of TBP. Further regulation is provided by SAGA-TBP interaction, principally through Spt8, in conjunction with Spt3 and/or other factors. The large white and black circles represent unidentified members of SAGA; several Taf_IIs are also present in SAGA (25), but their exact role in the complex has yet to be defined.