(A) The Rtt109 interface with the Vps75 earmuff domain is shown in cartoon representation. Residues that mediate protein-protein interaction or near the protein-protein interface and targeted for mutagenesis are shown as stick models in CPK coloring. Hydrogen bonds are indicated as yellow dashed lines. The orientation is similar to figure 1A (boxed interface 1).
(B) The Rtt109 interface with the Vps75 dimerization domain rendered as described above.
(C) A superposition of the Rtt109 active site in the apo- and Vps75-bound forms is shown in cartoon representation. Active site residues in the insert are shown as stick models in CPK coloring (Lys290 is too far to be included in the insert).

(A–B) Orthogonal surface representations of the Rtt109-AcCoA/Vps75 complex reveal a high degree of sequence conservation within the interior of the ring-shaped complex. The color-coding indicates the degree of conservation as shown in the bar key, with sequence conservation information derived from previously published alignments (Tang et al., 2008a; Tang et al., 2008b). For clarity, only one Rtt109 molecule and half of the Vps75 homodimer are shown (gray patch represents the cross section at point of omission). The orientation of (A) is the same as in Figure S2A.
(C) The exterior surface of the ring-shaped Rtt109/Vps75 complex shows a low degree of sequence conservation.
(D) Location of Rtt109 (Arg194 and Glu368) and Vps75 (Asp198, Glu206, and Glu207) residues whose mutation decreases HAT activity (panel E). These residues are not involved in mediating Rtt109-Vps75 interaction and do not appear to be part of the active site and, therefore, are suggested to participate in histone binding. The view is as in panel A.
(E) Summary of kinetic parameters of selected Rtt109 and Vps75 mutants within the ring interior of the Rtt109-AcCoA/Vps75 complex (Figure S4). All parameters were determined by a non-linear regression fit to the classical Michaelis- Menten equation (see methods). N.D. indicates not determined due to experimental considerations.

(A–B) Effect of selected Rtt109 mutations on the Rtt109-Vps75 interaction. Co-purification of Rtt109 and Vps75 from strains expressing Rtt109 mutants was monitored by immunoblotting (A) or mass spectrometry (B, see Supplemental Methods). All the strains used in (A) contain a wild-type chromosomal VPS75 gene and an additional plasmid encoding a wild-type VPS75-HA gene.
(C) Effect of selected Vps75 mutations on the Rtt109-Vps75 interaction. Co-purification of Rtt109 and Vps75 from strains expressing Vps75 mutants was monitored by immunoblotting.
(D) Stability of Rtt109 mutants. The stability of Rtt109 mutants was determined by immunoblotting as a function of time after the addition of cycloheximide. The top panel is a control showing the reduced stability of wild-type Rtt109 in cells that lack Vps75.

(A and C) Two orthogonal views of the Rtt109/Vps75 complex. The obligate Vps75 homodimer is shown with subunits in blue and green and the two Rtt109 subunits are shown in purple and orange. The proteins are shown in cartoon representation. The two interfaces that are boxed are shown in molecular detail in figures 2A and 2B.
(B and D) The same views as (A) and (C), respectively, but shown in surface representation.

(A) Pull-down studies of Vps75-wt and mutants by GST-Rtt109-wt. Published mutants (**) were included for comparison (Tang et al., 2008b). Vps75-(monomeric)# is Vps5-(C21E,V25S,I28E,V32E) (Berndsen et al., 2008).
(B) HAT assays were performed with (H3–H4)₂ tetramer and [¹⁴C]AcCoA substrates and complexes of GST-Rtt109-wt with Vps75-wt and mutants.
(C) Pull-down studies of Vps75-wt by GST-Rtt109-wt and mutants. A published mutant (**) was included for comparison (Tang et al., 2008b).
(D) Enzymatic studies of GST-Rtt109-wt and mutants in the presence of Vps75-wt or yAsf1N (1–154). HAT assays were performed with (H3–H4)₂ tetramer and [¹⁴C]AcCoA substrates.
(E) Limited proteolysis of Rtt109/Vps75 complexes containing wild-type, mutant or truncated forms of Rtt109. Reaction products after 10 min of trypsin digestion are shown. A 30 min reaction showed a nearly identical pattern of digestion.
(F) Catalytic rates (k_cat) were measured for Rtt109-WT and Rtt109-(R292E) using the following substrates in the context of the (H3–H4)₂ tetramer: H3-WT, H3-(K9R/K56R), H3-(K27/K56R), H3-(K9/K27R) and H3-(K9R, K27/K56R) (Figure S3). All parameters were determined by a non-linear regression fit to the Michaelis-Menten equation (see methods).

(A–F) Histone lysine specificity of gcn5, rtt109 and vps75 null mutants and selected mutants that specifically target the integrity of the Rtt109/Vps75 holoenzyme as determined by mass spectrometry. Total histones were purified from asynchronously growing cells and analyzed by mass spectrometry (Supplemental Methods) to determine the fraction of H3 molecules acetylated at K9, K27 and K56.
(G–H) Histone lysine specificity of selected Rtt109 and Vps75 mutants as determined by immunoblotting. 1.5-fold serial dilutions of whole-cell lysates (from left to right) were probed by immunoblotting to determine the levels of H3K9, K27 and K56 acetylation in cells expressing Rtt109 or Vps75 mutants. The asterisk points to a non-specific band that is weakly detected by our anti-H3K9ac in H3K9R mutant cells (not shown).
(I) Effect of selected Rtt109 mutants on histone assembly. Complexes of CAF-1 and histones H3/H4 were affinity-purified from yeast cells to determine the amounts of histones bound to CAF-1 in strains expressing Rtt109 mutants.