Sds3 (Suppressor of Defective Silencing 3) Is an Integral Component of the Yeast Sin3·Rpd3 Histone Deacetylase Complex and Is Required for Histone Deacetylase Activity*

SDS3 (suppressor of defective silencing 3) was originally identified in a screen for mutations that cause increased silencing of a crippled HMR silencer in arap1 mutant background. In addition, sds3mutants have phenotypes very similar to those seen in sin3and rpd3 mutants, suggesting that it functions in the same genetic pathway. In this manuscript we demonstrate that Sds3p is an integral subunit of a previously identified high molecular weight Rpd3p·Sin3p containing yeast histone deacetylase complex. By analyzing an sds3Δ strain we show that, in the absence of Sds3p, Sin3p can be chromatographically separated from Rpd3p, indicating that Sds3p promotes the integrity of the complex. Moreover, the remaining Rpd3p complex in the sds3Δ strain had little or no histone deacetylase activity. Thus, Sds3p plays important roles in the integrity and catalytic activity of the Rpd3p·Sin3p complex.

SDS3 (suppressor of defective silencing 3) was originally identified in a screen for mutations that cause increased silencing of a crippled HMR silencer in a rap1 mutant background. In addition, sds3 mutants have phenotypes very similar to those seen in sin3 and rpd3 mutants, suggesting that it functions in the same genetic pathway. In this manuscript we demonstrate that Sds3p is an integral subunit of a previously identified high molecular weight Rpd3p⅐Sin3p containing yeast histone deacetylase complex. By analyzing an sds3⌬ strain we show that, in the absence of Sds3p, Sin3p can be chromatographically separated from Rpd3p, indicating that Sds3p promotes the integrity of the complex. Moreover, the remaining Rpd3p complex in the sds3⌬ strain had little or no histone deacetylase activity. Thus, Sds3p plays important roles in the integrity and catalytic activity of the Rpd3p⅐Sin3p complex.
To counteract the effect of HAT complexes, it is necessary to reverse acetylation by efficiently deacetylating nucleosomal histones. Histone deacetylase complexes (HDACs) perform this reaction. HDACs have been isolated and characterized in several organisms as multiprotein complexes that are associated with DNA binding repressors and co-repressors (14 -22). The majority of these complexes contain members of the Rpd3⅐ HDAC-related protein family as catalytic subunits (23). In addition to these Rpd3p-related HDAC complexes, two non-Rpd3prelated deacetylase complexes have been identified in Zea mays (24,25). Moreover, yeast HDAC complexes containing Hda1p and Hos3p as catalytic subunits have been identified (26,27).
Two yeast multiprotein HDAC complexes have been found to contain Rpd3p (26,28). The larger of these complexes was also found to contain the Sin3p co-repressor (28). Sin3p is thought to mediate interactions of the Sin3⅐Rpd3 complex with sequence-specific DNA binding repressors such as Ume6p, which leads to a localized deacetylation of histones H3 and H4 and repression of transcription in vivo (19,29). RPD3 and SIN3 were both originally identified in genetic screens as regulators of gene expression (30,31). Mutations in RPD3 and SIN3 affect transcription of the same set of genes (32). SIN3 and RPD3 were identified among at least 19 other genes in a screen for extragenic suppressors of a silencing defective RAP1 allele (rap 1-12) (33,34). Another gene identified in this screen is SDS3 (suppressor of defective silencing 3). Although mutations in SDS3 were shown to cause several phenotypes in common with sin3 and rpd3 mutants, they did not appear to derepress a plasmid-borne TRK2 gene, raising the possibility that SDS3 might function independently of SIN3 and RPD3 (33). However, a recent study found that a mutation in SDS3 reduced Sin3p-mediated repression and that Sds3p and Sin3p could be co-immunoprecipitated from cell extracts (35). These results illustrated that SDS3 functions in the same genetic pathway as SIN3 and that Sds3p and Sin3p can interact in some way.
In this manuscript we demonstrate that Sds3p is an integral subunit of a Rpd3p⅐Sin3p-containing yeast HDAC complex with an apparent molecular mass of 1.2 MDa. By analyzing a sds3⌬ strain we show that, in the absence of Sds3p, Sin3p can be chromatographically separated from Rpd3p, indicating that Sds3p promotes the integrity of this complex. In addition, the remaining Rpd3p complex in the sds3⌬ strain had little or no histone deacetylase activity. Thus, Sds3p plays important roles in the integrity and catalytic activity of the Rpd3p⅐Sin3p complex.
* This work was supported in part by a grant from the National Institute of General Medical Sciences (to J. L. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Yeast Strains-The sds3⌬ strain in the W303 background (36) was generated as described previously (33). The Sds3-HAp construct was made by inserting three copies of the HA epitope in the 3Ј-end of the SDS3 open reading frame. A unique NotI site was engineered into ϩ978 of the 984-bp SDS3 open reading frame. A NotI cassette containing three copies of the HA epitope (YPYDPDYA) was inserted and cloned into the 2-m plasmid pRS423. The resulting clone, DV-246, complements the sds3⌬ mutation in a manner identical to the untagged SDS3 gene.
Preparation of Whole Cell Extracts and Purification of the Rpd3p Complex-Strains were grown to an optical density of 1.5 as described previously (37), and cells of a 12-liter culture were harvested by centrifugation at 3000 ϫ g for 10 min. The cell pellets were washed, resuspended, and lysed by using a glass bead-beater (Biospec), and the resulting extracts were loaded onto Ni 2ϩ -agarose as described (37). After Ni 2ϩ -agarose chromatography, the imidazole eluate was loaded directly onto a Mono-Q HR5/5 column (Amersham Pharmacia Biotech) to separate HDAC complexes. Bound proteins were eluted with a linear 25-ml gradient of 100 to 500 mM NaCl in buffer B (50 mM Tris-HCl, pH 8.0, 10% glycerol, 0.1% Tween 20, 2 g/ml leupeptin, 2 g/ml pepstatin A, 1 mM PMSF, 0.5 mM DTT). Fractions of 0.5 ml were collected and assayed for HDAC activity as described previously (38). In general, 10 l of each fraction was incubated with 3 g of tritium-labeled chicken reticulocyte core histones (38). Samples were incubated for 30 min at 30°C, and the released radioactivity was measured as described (38). Fractions containing HDAC activity were pooled separately, concentrated to 500 l using Centriprep 10 (Millipore) concentrators, and then loaded onto a Superose 6 HR10/30 size exclusion column (Amersham Pharmacia Biotech) to determine the native molecular weight. The column was run in 350 mM NaCl in 40 mM Hepes, pH 7.8, 10% glycerol, 0.1% Tween 20, 2 g/ml leupeptin, 2 g/ml pepstatin A, 1 mM PMSF, 0.5 mM DTT at a flow rate of 0.2 ml/min. Fractions of 0.5-ml volume were collected and assayed for HDAC activity. Aliquots of fractions containing HDAC activity were applied to SDS-PAGE and subject to Western blotting as described (39). In addition, 50 l of Mono-Q concentrate was applied to a Superose 6 PC 3.2/30 size exclusion column (Amersham Pharmacia Biotech). The column was run in 500 mM NaCl in 40 mM Hepes, pH 7.8, 10% glycerol, 0.1% Tween 20, 2 g/ml leupeptin, 2 g/ml pepstatin A, 1 mM PMSF, 0.5 mM DTT at a flow rate of 0.02 ml/min. 50-l fractions were collected and tested for HDAC activity, and Western blot analysis was performed as described. The flow-through fraction of Ni 2ϩ -agarose chromatography was dialyzed three times against 10 volumes of buffer B (100 mM NaCl, 50 mM Tris-HCl pH 8.0, 10% glycerol, 0.1% Tween 20, 2 g/ml leupeptin, 2 g/ml pepstatin A, 1 mM PMSF, 0.5 mM DTT) and processed over a Mono-Q column and a Superose 6 size exclusion column. Western blot analyses described in this study were performed using antibodies against Rpd3p (Upstate Biotechnology), Hda1p (Santa Cruz Biotechnologies), HA (Covance), and Sin3p (28).
Immunoprecipitation, Modified HDAC Assay for Immunoprecipitations-Antibodies for HA (Covance) and Rpd3p (Upstate Biotechnology) were coupled to protein A-Sepharose beads. 40 l of 50% bead slurry was washed twice with 200 l of binding buffer (150 mM NaCl, 40 mM Hepes, pH 7.5, 10% glycerol, 0.1% Tween 20, 2 g/ml pepstatin A, 2 g/ml leupeptin, 0.2 mM DTT, and 1 mM PMSF). Beads were recovered by centrifugation at 110 ϫ g for 1 min in a table-top centrifuge after each wash step. 20 l of antibody solution (20 g) was added to the beads, and the antibodies were bound to the beads by rotation on a wheel for 1 h at room temperature. Antibodies were cross-linked to beads as described previously (39). Superose 6 chromatography peak fractions were diluted to the appropriate salt concentration of 150 mM NaCl with binding buffer, reconcentrated with Microcon 10 concentrators (Amicon) to the original volume, and 20 l of sample was added to 20 l of antibody beads. The binding reaction was performed for 4 h or overnight with a rotation wheel at 4°C. After the binding reaction, beads were recollected by centrifugation, the supernatant was removed and saved, and beads were washed twice with binding buffer. HDAC assays were performed with beads and supernatant. 20 l of 1 g/l radiolabeled chicken erythrocyte core histones were added to samples and were incubated for 3 h at 30°C with rotation. HDAC activity was measured as described previously (38).
Whole cell extract for immunoprecipitation was prepared from 100 ml of the Sds3-HAp expression strain by glass bead disruption into extraction buffer (40 mM HEPES, pH 7.5, 350 mM NaCl, 10% glycerol, 0.1% Tween 20, 1 mM PMSF, 1 mM DTT, 2 g/ml pepstatin A, 2 g/ml leupeptin). For immunoprecipitation, 200 g of whole cell extract total protein, as determined by Bradford assay, was diluted to 150 mM NaCl in extraction buffer lacking NaCl. The diluted extract was mixed for several hours at 4°C in the presence or absence of 4 g of antibody directed against Rpd3 or 1 l of antibody directed against Sin3. Immune complexes were collected by mixing antibody-or mock-treated extracts with protein A-Sepharose for several hours at 4°C. After several washes in binding buffer (see above), input extract, unbound supernatants, and bound bead material were subjected to Western analysis using antibody against the HA epitope tag.
For Western blot analysis, beads and supernatant were boiled in Laemmli SDS sample buffer for 10 min and applied to 10% SDS-PAGE with subsequent Western blotting as described (39).

Identification of Yeast HDAC Complexes-Several yeast
HAT complexes bind to Ni 2ϩ -agarose, which concentrates these activities from yeast whole cell extract for further purification (37). We tested if HDAC activities might also be found in the Ni 2ϩ -agarose eluant to determine if this material might also serve as a starting point for purifying HDAC multiprotein complexes. Whole yeast cell extract of a Sds3-HAp strain was prepared and bound to Ni 2ϩ -agarose. The flow-through and eluates from Ni 2ϩ -agarose were then subjected to Mono-Q chromatography followed by Superose 6 size exclusion chromatography ( Fig. 1). We detected three HDAC activities by this procedure. These included an Rpd3-containing complex of approximately 0.6 MDa (data not shown), which was found in the Ni 2ϩ -agarose flow-through and is presumably related to that described by Rundlett and co-workers (26). We did not observe co-fractionation of Sds3-HAp with this Rpd3 complex (data not shown). Two additional HDAC complexes eluted together from the Ni 2ϩ -agarose but were separated on the subsequent Mono-Q column ( Fig. 2A). Using Western blot analysis we found that Rpd3p, Sin3p, and Sds3-HAp co-fractionated with the HDAC complex eluting at approximately 0.35 M NaCl ( Fig.  2A). A second HDAC activity eluted at approximately 0.25 M NaCl, and subsequent Western blot analysis indicated that it co-elutes with Hda1p ( Fig. 2A).
Co-fractionation of Sds3p, Rpd3p, and Sin3p-To further test if Sds3p is an actual component of the Rpd3p⅐HDAC complex, we tested for co-elution of Sds3-HAp with Rpd3p, Sin3p, and HDAC activity by gel filtration chromatography. Mono-Q fractions containing the Rpd3p complex from the Sds3-HA strain (fractions 28 -34) were pooled, concentrated, and applied to a Superose 6 size exclusion column. The HDAC activity eluted at an apparent molecular mass of 1.2 MDa (Fig. 2B). This was consistent with our earlier findings showing the same molecular weight for this particular HDAC complex from a wild type strain (data not shown). Aliquots of fractions 16 -24 of Superose 6 size exclusion chromatography were applied to SDS-PAGE. Subsequent Western blotting with antibodies for Rpd3p, Sin3p, and HA (Sds3p) demonstrated co-elution of all three of these proteins with the deacetylase activity peak (Fig.  2B). The high molecular weight of this complex and the fact that it contains Sin3p and Rpd3p suggest that it is related to or identical to the complex described by Kasten and colleagues (28).
Co-immunoprecipitation of Sds3p, Rpd3p, Sin3p, and HDAC Activity-To confirm that Sds3p is a bona fide subunit of the 1.2-MDa Rpd3p complex, we tested for co-immunoprecipitation of these two proteins and HDAC activity from the Superose fractions. We used fraction 19 of the Superose 6 size exclusion chromatography of the Sds3-HA strain, which corresponds to the HDAC activity peak (Fig. 2B) and performed co-immunoprecipitation experiments. Samples were incubated with HA antibodies coupled to beads, the beads were pelleted, and the HDAC activity of supernatant and beads was tested. Fig. 3A shows that the HDAC activity was clearly immunoprecipitated with the HA antibodies, illustrating that Sds3p is part of the HDAC complex. As a control we incubated samples of a HDAC peak fraction of a Superose 6 size exclusion chromatography of a SDS3 ϩ (WT) strain with HA antibodies coupled to beads. Fig.  3B (group 1) shows that HDAC activity was not immunoprecipitated with the HA antibodies from this strain, ruling out any nonspecific interaction of HA antibodies with the Rpd3p complex. Moreover, the HDAC activity of the Rpd3p complex from both the Sds3-HA and wild type strains immunoprecipitated with Rpd3p antibodies coupled to beads (Fig. 3B, groups  2 and 3). To further confirm the co-immunoprecipitation of Sds3p and Rpd3p, Western blots of the immunoprecipitations with anti-HA were performed. In Fig. 3C (upper panel), Rpd3p and HA-Sds3p can be detected in the bead fraction using fraction 19 of the Superose 6 column of the Sds3p-expressing strain, whereas no signal is detectable in the supernatant. Lanes 2 and 3 of Fig. 3C (lower panel) show co-immunoprecipitation of Rpd3p with antibody against Rpd3 using Superose 6 peak fraction of the WT, untagged strain. By contrast, no Rpd3p protein was immunoprecipitated with antibodies for HA from this strain (Fig. 3C, lower panel, lanes 4 and 5).
Further confirmation that Sds3 associates with the Rpd3⅐Sin3 complex came from immunoprecipitation experiments performed in crude extracts. Antibodies directed against either Rpd3 or Sin3 immunoprecipitated Sds3 from whole cell extract (Fig. 3D) (35). As expected, antibody directed against Rpd3 was also able to immunoprecipitate Sin3 from the same extract (data not shown). These results indicate that, in addition to highly fractionated preparations, Rpd3, Sin3, and Sds3 associate within cell extracts.
Deletion of SDS3 Alters the Chromatographic Behavior of the Rpd3⅐Sin3 Complex-To investigate if the integrity of the 1.2-MDa HDAC complex was dependent on SDS3, we prepared whole cell extracts from a sds3⌬ strain and applied it to Ni 2ϩagarose followed by Mono-Q chromatography. The HDAC activity in the fractions normally containing the WT 1. The HDAC Activity of the Rpd3 Complex Depends on SDS3-To determine if the deletion of SDS3 affects not only the elution profile of Sin3p and Rpd3p, but also the integrity and activity of the complex, we pooled fractions 20 -32 of the Sds3⌬ Mono-Q chromatography, concentrated the material, and applied it to Superose 6 size exclusion chromatography. The HDAC activity was examined as described, and corresponding aliquots were analyzed by SDS-PAGE and subsequent Western blotting. Fig. 4B shows that the main HDAC activity eluted at approximately 0.6 MDa and corresponds to Hda1p, indicating that the deletion of SDS3 does not affect a 0.6-MDa Hda1p complex. However, the remaining Rpd3 com- plex had a smaller molecular weight compared with wild type complex (Fig. 2B). Western blot analysis revealed a signal for Rpd3 at fractions 20 and 21, which corresponds to an approximate molecular mass of 0.9 MDa. Importantly, this 0.9-MDa Rpd3 breakdown did not have HDAC activity.
Deletion of SDS3 Decreases the Interaction between Sin3p and Rpd3p-The elution of Sin3p from the Superose 6 column (Fig. 4B) overlapped with that of Rpd3p. However, these pro- FIG. 4. HDAC activity profile of the sds3⌬ strain. A, whole cell extract of a sds3⌬ strain was prepared and applied to Ni 2ϩ -agarose and subsequently to a Mono-Q column. The HDAC elution profile shows one broad HDAC peak, which corresponds to the Hda1p complex found from wt cells (Fig. 2). Western blotting revealed the elution of Hda1p, Sin3p, and Rpd3p from the sds3⌬ strain (compare with Fig. 2A). B, the Superose 6 elution profiles from the sds3⌬ strain. Fractions 20 -32 of Mono-Q of the sds3⌬ strain (A) were pooled and applied to Superose 6 size exclusion chromatography at 0.35 M NaCl (B). The HDAC activity eluted at approximately 0.6 MDa with Hda1p. No HDAC activity was eluted with Sin3p or Rpd3p from this strain.
in the bead-bound fraction immunoprecipitated with antibody for HA (lane 5). D, co-immunoprecipitation of Sds3p with Rpd3p and Sin3p from whole cell extracts. Whole cell extract from a Sds3-HAp expressing strain was subjected to immunoprecipitation conditions in the presence or absence of antibody directed against Rpd3 or Sin3. The bead-bound material (B) and one-fifth of the input extract and unbound supernatants (S) were subjected to Western analysis for Sds3-HAp using antibody recognizing the HA-epitope tag. FIG. 3. Co-immunoprecipitation of Sds3p, Rpd3p, Sin3p, and HDAC activity. A, the peak fraction, fraction 19 of the Superose 6 column from the Sds3-HAp-expressing strain ( Fig. 2A) was used to perform immunoprecipitation experiments. The HDAC activity of this fraction was found to co-immunoprecipitates with Sds3-HAp. B, immunoprecipitation of the HDAC activity with antibody for HA is specific for the Sds3-HAp-expressing strain. Superose 6 HDAC peak fraction of the 1.2-MDa complex prepared from an SDS3 ϩ strain (WT) was incubated with antibody for Hap (group 1). Superose 6 HDAC peak fractions from the Sds3-HA-expressing strain and an SDS3 ϩ strain (Sds3-HA and WT) were precipitated with antibody for Rpd3p (groups 2 and 3, respectively). Input (I), supernatant (S), and bead-bound fraction (B) were monitored for HDAC activity in all three cases. HA antibody does not precipitate HDAC activity of WT strain (group 1), whereas Rpd3p antibody precipitates HDAC activity of Sds3-HA and WT strain (groups 2 and 3). C, the Western blot analysis of these immunoprecipitation experiments is shown. Top panel, the Superose 6 peak fraction from the Sds3-HA strain (Input, see Fig. 2B) was incubated with HA beads. Input (I), supernatant (S), and bead-bound fraction (B) were separated on SDS-PAGE followed by Western blotting. Blots were probed with antibodies to HAp and Rpd3p. Bottom panel, the Superose 6 peak fraction from the SDS3 ϩ strain (Input) was incubated with ␣Rpd3 beads and ␣HA beads and further processed as described above. Rpd3 in the Superose 6 peak fraction from the WT strain immunoprecipitates with antibody to Rpd3p (lane 3). By contrast, no Rpd3p signal can be observed teins did not co-elute to the extent that they did from the wild type strain (Fig. 2B). This suggests that the association of Sin3p and Rpd3p was weakened in the absence of Sds3p. To confirm that Sin3p and Rpd3p could be chromatographically separated in the sds3⌬ strain, we applied the concentrated Mono-Q pool (fractions 20 -32, see Fig. 4A) to an additional Superose 6 size exclusion column, using the Amersham Pharmacia Biotech SMART (Sensitive Methods And Recovery Technology) system, at 0.5 M NaCl. HDAC activity was measured, and Western blot analysis was performed (Fig. 5A). The remaining HDAC activity eluted at 0.6 MDa and corresponded to Hda1p. Rpd3p was detected in fractions 24 and 25, which correspond to a molecular mass of approximately 0.9 MDa. Sin3p was now detected in the molecular range of 0.7-0.5 MDa. Thus, under these chromatographic conditions, Rpd3p and Sin3p were separated into distinct subcomplexes in the absence of Sds3p (Fig. 5A). As a control, the concentrated Mono-Q pool of the Sds3-HA strain (fractions 20 -32, see Fig. 2A) was applied to Superose 6 on the SMART system under the same conditions. Fig. 5B shows that two HDAC activities were separated. One activity corresponded to a 1.2-MDa complex, containing Sin3p, Rpd3p, and Sds3p, whereas the second activity, eluting at 0.6 MDa, corresponded to the Hda1p complex. This confirmed that the higher salt concentration used on this column did not cause a disruption of the WT Rpd3p⅐Sin3p⅐Sds3p complex.

DISCUSSION
To understanding the detailed roles of Rpd3p⅐HDAC complexes, it will be necessary to identify the functions of other subunits of these multiprotein complexes. It was shown previously that Sin3p targets Rpd3p-dependent HDAC activity through interaction with the repressor protein Ume6 to certain promoters (19,29). Therefore, it is very likely that other subunits are required for the regulation of the enzymatic activity of HDAC complexes. Sap 30, for example, was identified as part of Sin3p⅐HDAC and Rpd3p⅐HDAC complexes in mammals and yeast (20,40). It appears to be required for the normal function of these complexes. Furthermore, Sap30 is capable of repressing transcription when tethered to DNA. This might indicate that Sap30 facilitates interactions of Sin3p⅐HDAC and Rpd3p⅐HDAC and might even recruit HDAC activity in the absence of Sin3p (20). Disruption of SAP30 in yeast shows phenotypes comparable to disruption of SIN3 and RPD3, suggesting that it works in the same genetic pathway. Moreover, Sap30p and Rpd3p have been shown to co-immunoprecipitate with antibodies against Rpd3p (40).
Deletion of another yeast gene, SDS3, also showed similar phenotypes to deletions of SIN3 and RPD3. The yeast SDS3 gene was originally identified in a screen for mutations that cause increased silencing of a crippled HMR silencer in a rap1 mutant background (33,41). This screen identified more than 20 other genes, including SIN3 and RPD3. Subsequent analysis showed that SDS3 shares several transcription regulation properties with SIN3/RPD3, although epistasis tests of the silencing effect suggested that SDS3 might differ in function, at least subtly, from RPD3/SIN3 (33). However, more recent work has shown that sds3 mutants have phenotypes very similar to those seen in sin3 and rpd3 mutants, with transcriptional regulation of the same set of genes affected in all three mutants (35). The changes in transcriptional regulation seen in rpd3/sds3 and sin3/rpd3 double mutants are no more severe than the single mutants. This genetic analysis indicates that SDS3 is in the same functional pathway as RPD3 and SIN3, and this idea is supported by co-immunoprecipitation experiments showing that Sds3p can associate with Sin3p (35).
To determine whether Sds3p is part of the Rpd3p⅐Sin3p complex, we reintroduced an HA-tagged SDS3 gene into the sds3⌬ strain. The reintroduced gene rescued the complex. We used this strain to partially purify the complex and used Western blot analysis to illustrate co-elution of Sin3p, Rpd3p, and Sds3p. Co-immunoprecipitation experiments using material obtained from the Sds3-HAp strain demonstrated that Sds3p is an integral part of the 1.2-MDa Rpd3p⅐Sin3p complex. We were able to immunoprecipitate Rpd3p and HDAC activity with antibodies against HA-tagged Sds3p. Deletion of SDS3 had a significant effect on the Rpd3p⅐Sin3p complex. First, the size of the Rpd3p portion of the complex was reduced from approximately 1.2 to 0.9 MDa. Second, the association of Sin3p with the Rpd3p complex was weakened. Sin3p was chromatographically separable from Rpd3p in the absence of Sds3p. This result suggests that Sds3p promotes the interactions of Sin3p with Rpd3p or other components of the complex. Sds3p may not, however, be absolutely required for the interaction of  4A) were applied to Superose 6 on the SMART system at 0.5 M NaCl. Rpd3 eluted at fractions 23-24, corresponding to an approximate molecular mass of 0.9 MDa. No HDAC activity was detected in this molecular range. Sin3p eluted in the range of 0.7-0.5 MDa. Hda1p eluted at approximately 0.6 MDa, which corresponds to the measured HDAC activity (A). B, as a control for A, the corresponding fractions from the Sds3-HAp-expressing strain were analyzed on the Superose 6 column at 0.5 M NaCl using the SMART system. Mono-Q fractions 20 -36 (Fig. 2B) were applied to Superose 6 on the SMART system. Two HDAC peaks were separated corresponding to the 1.2-MDa Sin3p, Rpd3p, and Sds3-HAp complex and the 0.6-MDa Hda1p complex, respectively.
Sin3p with the Rpd3p complex. At low salt concentration some Sin3p and Rpd3p co-elute in the absence of Sds3p and they can be co-immunoprecipitated from extract made from an SDS3 deletion strain. 2 Finally, in the absence of Sds3p the Rpd3p complex lacked HDAC activity. Thus, the association of Sds3p, Sin3p, or another subunit that might dissociate in the absence of Sds3p is required for the ability of Rpd3p to act as a histone deacetylase. It is important to note, however, that the second Rpd3p-containing complex (0.6 MDa) identified in this study, which does not contain Sds3p, is fully active on histone substrates. Thus, in a different context, the HDAC activity of Rpd3 apparently functions in the absence of Sds3p. Alternatively, a different protein in the smaller complex may replace the function of Sds3p.
In conclusion, our observations indicate that Sds3p is necessary to maintain the function of a yeast Sin3p⅐Rpd3p complex. Sds3p facilitates protein-protein interactions within the complex to maintain its structure and enzymatic activity. Sds3p might provide part of the linkage between the co-repressor region of the complex (Sin3p) and the deacetylase function (Rpd3p). Furthermore, disruption of SDS3 abrogates the ability of the 1.2-MDa Rpd3⅐Sin3 complex to function as a histone deacetylase.