A novel role for Sem1 and TREX-2 in transcription involves their impact on recruitment and H2B deubiquitylation activity of SAGA

Yeast strains, DNA recombinant work and microbiological techniques

The yeast strains used in this study are listed in Supplementary Table S1, along with the quantitative polymerase chain reaction (qPCR) primers and antibodies (Supplementary Tables S2 and S3). Microbiological techniques and yeast plasmid transformation were essentially done as described previously (7). The chromosomal integration of TAP (URA3 marker), MYC (HIS3 marker) and C-terminal tags was performed as previously described (28,29). For gene disruptions, the indicated gene was deleted by high-efficiency transformation using a PCR product amplified from either the KanMX4 plasmid pRS400 or the HIS3 plasmid pFA6a. All the deletions and genomically tagged strains were confirmed by PCR analysis and/or western blot analysis. Strains were grown under standard conditions. For the growth analysis, yeast cells were diluted to 0.5 OD₆₀₀, and serial dilutions (1:10) were spotted onto YP + glucose and incubated at various temperatures.

TAP purifications, immunoprecipitations and western blot analysis

Purification of Sus1-tandem affinity purification (TAP), Ada2-TAP and Ubp8-TAP in wild-type (BY4741) and mutant strains was performed as previously described (7). TAP fusion proteins and associated proteins were recovered from cell extracts by affinity selection in an IgG matrix. After washing, the Tobacco Etch Virus (TEV) protease was added to release the bound material. The eluate was incubated with calmodulin-coated beads in the presence of calcium. This second affinity step was required to remove not only the TEV protease but also traces of contaminants remaining after first affinity selection. After washing, the bound material was released with ethylene glycol tetraacetic acid (EGTA). This enriched fraction was called calmoduline eluate. The calmoduline eluates from the TAP-purified complexes were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). The western analysis of calmoduline eluates was performed using anti-MYC, anti-Sem1 polyclonal or anti-TAP according to standard procedures.

To analyze the TAP-purified protein complexes, trichloroacetic acid (TCA)-precipitation, LysC/trypsin digestion and multidimensional protein identification technology (MudPIT) analyses by mass spectrometry were performed as described elsewhere (30). The immunoprecipitation of Sus1-MYC, Ubp8-TAP and Ada2-TAP in wild-type and mutants was performed as follows: 50 ml of yeast cultures were grown on rich medium to a 0.5 OD₆₀₀. Cells were harvested, washed with water and resuspended in 250 μl of lysis buffer [50 mM HEPES–KOH at pH 7.5, 140 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 10% glycerol, 0.5% NP-40, 1 mM PMSF and protease inhibitors]. An equal volume of glass beads was added. Breakage was achieved by four pulses of vortexing at 4°C lasting 1 min. The clarified extracts were immunoprecipitated for 60 min at 4°C using anti-MYC antibodies or IgG Sepharose six Fast Flow resin (GE Healthcare). Immunoprecipitates were washed three times with 1 ml of lysis buffer for 10 min and were subsequently resuspended in 50 μl of SDS–PAGE sample buffer. Western analysis was performed using anti-MYC, anti-Sem1 and anti-TAP according to standard procedures.

Yeast whole-cell extracts (WCEs) were prepared as described elsewhere. Typically, 7 μl of WCE were loaded on SDS–PAGE gels. After electrophoresis and transfer to nitrocellulose membranes, samples were subjected to immunoblot analyses using procedures and detection reagents from ECL Amersham®.

Chromatin immunoprecipitation and double chromatin immunoprecipitation

Chromatin immunoprecipitation (ChIP) and chromatin double immunoprecipitation (ChDIP) were performed as formerly described (31,32) using 100 ml (ChDIP) or 50 ml (ChIP) of early-log-phase cultures grown in YP + glucose (repressing), YP + raffinose (de-repressing) or YP + galactose (inducing) for inductions of 25 (ChIP) or 90 min (ChDIP). Cultures were cross-linked with 1% of formaldehyde solution (Sigma) for 20 min at room temperature with intermittent shaking and were then quenched with 125 mM glycine. Cells were collected by centrifugation and washed four times with 25 ml cold Tris–saline buffer (150 mM NaCl and 20 mM Tris–HCl at pH 7.5). Cell breakage was performed in 300 ml of lysis buffer (50 mM HEPES–KOH at pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% Tergitol-type NP-40 (NP-40), 1 mM phenylmethanesulfonylfluoride or phenylmethylsulfonyl fluoride (PMSF) and protease inhibitors) with glass beads, and cell extracts were sonicated in a Bioruptor sonicator (Diagenode) for 30 min in 30 s on/30 s off cycles (chromatin was sheared into an average size of 300 bp). Ten microliters of extract was reserved as the input. The rest was used for immunoprecipitation with magnetic beads (Dynabeads©) coated with polyclonal goat anti-mouse IgG antibodies and was incubated O/N with 2.5 μl of mouse α-MYC antibodies (Sus1-MYC, Gcn5-MYC, Ada2-MYC, Spt8-MYC and Taf9-MYC in wild-type and mutants), or with magnetic beads coated with monoclonal human anti-mouse IgG antibodies (Rpb3-TAP) or by adding 90 μl of M2 agarose slurry (Sigma A2220; pre-washed three times in 1 ml of 125 mM glycine and equilibrated in lysis buffer for 60 min at 4°C) to extracts (wild-type and mutant sem1Δ). Immunoprecipitations were conducted overnight at 4°C, and the immune complexes were washed twice with lysis buffer, twice with lysis buffer supplemented with 360 mM NaCl, twice with wash buffer (10 mM Tris–HCl at pH 8.0, 250 mM LiCl, 0.5% NP-40, 125 mg of nadeoxycholol and 1 mM EDTA) and once with TE 1×. For ChDIP, Flag-H2B was eluted with 25 μl of 3× Flag peptide (Sigma F4799; 4 mg/ml of stock concentration) and incubated overnight at 4°C. Twenty microliters was separated as the input, and a second purification of Flag-H2B pool was performed using 2.5 μl of the anti-hyaluronic acid (HA) antibody (Roche 12CA5, 5 mg/ml), previously incubated with 50 μl of magnetic beads coated with polyclonal goat anti-mouse IgG antibodies (pre-washed three times in 1 ml of phosphate-buffered saline–bovine serum albumin and resuspended in 47.5 μl) and shaken overnight at 4°C. Immunoprecipitation was washed as previously described, and samples were eluted at 65°C for 15 min with 100 μl of elution buffer (50 mM Tris–HCl at pH 8, 10 mM EDTA and 1% SDS). Inputs and immunoprecipitation (IP) samples were incubated overnight at 65°C to reverse the cross-link. Samples were then treated with proteinase K (Ambion), at 100 μg/250 μl of chromatin for 2 h, and DNA was extracted twice with phenol:chloroform:isoamyl alcohol (25:24:1) and once with chloroform:isoamyl alcohol (24:1), ethanol precipitated and was resuspended in 40 μl TE. DNA was used for the qPCR reaction by using the appropriate primers listed in Supplementary Table S2. In all cases, at least three biological replicates were performed to obtain the standard errors of each experiment.

RNA isolation reverse transcriptase–PCR quantitative PCR

Total RNA was harvested from the yeast early-log-phase cultures grown in YP + glucose (repressing), YP + raffinose (de-repressing) or YP + galactose (inducing) by phenol/chloroform extraction (33). RNA was quantified, and quality was checked using Nanodrop®. The reverse transcriptase (RT)–PCR analysis was performed using 1 μg of total RNA. After DNase-I treatment, RNAs were purified by phenol/chloroform extraction. Reverse transcription was performed by following standard procedures with random hexamers and M-MLV reverse transcriptase (Invitrogen®).

Specific pairs of primers were used to amplify the qRT–PCR products of transcripts SCR1 or GAL1 using 3 μl of cDNA as a template (previously diluted at 1/10). For qPCR, SYBR® Green qPCR MM no ROX (Fermentas) was used in 10 μl of the final volume. Each sample was analyzed in triplicate, and qPCR was performed in a LightCycler® Roche 480. An activation step of 10 min at 95°C, followed by 40 cycles of 10 s at 94°C, 15 s at 57°C and 15 s at 72°C, was used for the GAL1 Open Reading Frame (ORF), ARG1 Promoter (PR) and SCR1 primer pairs. For the GAL1 PR pair, the Tm was set at 57°C. Real-time PCR amplification efficiency was calculated from the given slopes in the LightCycler® Roche 480. All the calculations were done by taking into account the real primer efficiencies. See Supplementary Table S2 for a list of these primers. The GAL1 and ARG1 mRNA fold-change (FC) levels were normalized to the reference gene SCR1 and were expressed in relation to the transcript level for each strain at the zero time point, defined as 1.0 (34,35). The ChDIP data were calculated as the signal ratio of IP samples in relation to the input signal minus the H2BK123R and H2BK123R sem1Δ mutant strain signal, set as a background control. The ChIP data were calculated as the signal ratio of IP samples in relation to the input signal minus the background of a no-antibody control. For the H3K79me3 signal, normalization was done in relation to total histone H3 levels. In all cases, at least three biological replicates were performed to obtain the standard errors of each experiment.

In vitro deubiquitylation assay

The Flag-tagged H2B substrate (containing ubH2B and unmodified H2B) was obtained by purifying N-terminally Flag-tagged histone H2B using an M2 agarose slurry (Sigma A2220). This substrate was quantified, and equal amounts (500 ng) were divided and incubated at room temperature for 30 min in deubiquitylation (DUB) buffer (100 mM Tris–HCl at pH 8.0, 1 mM EDTA, 1 mM DTT, 5% glycerol, 1 µM PMSF and 1% protease inhibitor) with the calmoduline eluate of each purification (Sus1-TAP and Ubp8-TAP) from wild-type (BY4741) and mutant strains (see Supplementary Figure S3A for the description of the experimental set-up). As a negative control, histone H2B was incubated solely in DUB buffer (which corresponds with the input of the assay) or with the calmoduline eluates of the Sus1-TAP sgf73Δ mutant, where the DUB activity associated with SAGA was impaired. The reaction was stopped by freezing on dry ice, followed by boiling in 1 volume of 2× SDS sample buffer for 2 min and running in a 15% SDS–PAGE gel. Gels were transferred to nitrocellulose, and a western blot analysis was performed using anti-HA antibodies to detect ubiquitin. The detection with this antibody allows us to identify specifically ubiquitylated histone H2B species, which makes possible accurate band quantification in each strain. Using anti-Flag antibodies allows us to identify all histone H2B species, taking into account that ubiquitylated histone H2B exists in low amounts within the cell; this is not a truthful method for ubH2B quantification (36). In all cases, at least three biological replicates were performed to obtain the standard errors of each experiment.

Fluorescence in situ hybridization

Fluorescence in situ hybridization against poly(A)⁺ RNA was done by growing yeast cells in 50 ml of YP + glucose medium at 30°C or 37°C at 0.5 OD₆₀₀. Cells were immediately fixed by adding 10% of formaldehyde for 60 min at room temperature. The fixative was removed by two rounds of centrifugation and was washed with 0.1 M potassium phosphate (pH 6.4). Cells were resuspended in ice-cold washing buffer (1.2 M sorbitol and 0.1 M potassium phosphate, pH 6.4), and, subsequently, the cell wall was digested with 0.5 mg/ml of zymolyase 100 T, whereas samples were applied on poly-l-lysine-coated slide wells. Non-adhering cells were removed by aspiration, cells were rehydrated with 2× SSC (0.15 M NaCl and 0.015 M sodium citrate) and were hybridized overnight at 37°C in 20 µl of pre-hybridization buffer (formamide 50%, dextran sulfate 10%, 125 µg/ml of Escherichia coli tRNA, 4× SSC, 1× Denhardt’s solution and 500 µg/ml of herring sperm DNA) with 0.8 pmol of Cy3-end-labeled oligo(dT) in a humid chamber. After hybridization, slides were washed with 1× SSC at room temperature, air-dried and mounted using VECTASHIELD® Mounting Medium with DAPI. Cy3-oligo(dT) was detected under a Leica DM600B fluorescence microscope. In all cases, at least three biological replicates were performed to obtain the standard errors of each experiment.

Image analysis

The nuclear/cytoplasmic ratio of fluorescence intensity was quantified using the Image J public domain Java image processing program (http://rsb.info.nih.gov/ij/download.htm). To determine the poly(A)⁺ RNA retention phenotype, the fluorescence intensity of the nucleus and cytoplasm in the wild-type and mutant strains images was measured, and the ratio was calculated by dividing the average nuclear intensity by the average cytoplasmic intensity. Data were then normalized to the mex67-5 control sample set as 100% of poly(A)⁺ RNA retention. A minimum of 100 cells per image were analyzed per strain (37). Error bars represent SE for at least three independent experiments.

DNA microarray expression profiling

Each mutant strain was profiled four times from two independently inoculated cultures and was harvested in early mid-log phase in synthetic complete medium with 2% glucose. Dual-channel 70mer oligonucleotide arrays were used, with a common reference wild-type RNA, hybridized in dye-swap. Wild-type cultures grown alongside mutants were processed in parallel and also hybridized in dye-swap versus the common reference wild-type RNA sample. After quality control, normalization and dye-bias correction (38), the replicate mutant profiles were averaged and statistically compared with a collection of 200 wild-type cultures, using Limma and Benjamini–Hochberg correction as described in Lenstra et al. (39). The reported FC is the average of the four replicate mutant profiles versus the average of all the wild-types. Shown in the figures are all genes that change at least once significantly (P < 0.05, FC > 1.7) in any single mutant.

Sem1 influences physical interaction of Sus1 with TREX-2

Sem1 is a component of both the proteasome (40) and the COP9 signalosome (41,42). It has also been found to interact with TREX-2 components Sac3 and Thp1 (42,43). More recently, the crystal structure of the Sac3–Thp1–Sem1 segment of the Saccharomyces cerevisiae TREX-2 complex has been solved (18). The small Sem1 protein wraps around the surface of Thp1 and makes few contacts with Sac3 (44). To investigate whether Sem1 interacts with Sus1, which happens for other TREX-2 subunits, Sus1 was TAP-purified, and the presence of Sem1 was analyzed. The western blot analysis using anti-Sem1 antibodies and the analysis of Sus1-TAP eluates by MudPIT reveal that Sem1 co-purifies with Sus1 via standard TAP purification (Figure 1A). To better characterize the Sus1–Sem1 interaction, its dependency on the presence of TREX-2 factors was also investigated. Figure 1B shows that both TREX-2 subunits Sac3 and Thp1 are required for the Sus1–Sem1 association. These results suggest that the interaction between Sus1 and Sem1 might be indirect and could be mediated by other TREX-2 components. Although considerable biochemical and structural information on the TREX-2 complex is available, the relevance of Sem1 in the Sus1 association with TREX-2 has not yet been investigated. Interestingly, although Sem1 and Sus1 have been found to interact with different regions of the complex, absence of Sem1 destabilizes the Sus1 association to TREX-2. Figure1C and D, respectively, illustrates how Sus1 binding to Sac3 and Thp1 is significantly reduced on Sem1 deletion. Thus, Sem1 influences the association of Sus1 with TREX-2, probably by stabilizing the complex or by promoting the assembly of the different subunits. The influence of Sem1 in Sus1 binding to TREX-2 prompted us to test whether Sem1 contributes to the role of Sus1 in mRNA export. Quantitation of the nuclear accumulation of poly(A)⁺ RNA in relation to cytoplasm concentration showed a 45% enrichment when SUS1 deletion was combined with SEM1 deletion (Figure 1E). Thus, the absence of SEM1 enhances the mRNA export defect provoked by SUS1 deletion, suggesting that they function independently to regulate mRNA export.

Sem1 interacts functionally with SAGA

Genetic and physical interactions between components of TREX-2 and SAGA in addition to Sus1 have been shown (45). For instance, mutants of SAC3 and THP1 are synthetic lethal with the deletion of SGF73 (46). To test whether SEM1 is also genetically related to SAGA, double mutants bearing SEM1 deletion were constructed in cells lacking GCN5, SGF73 and SUS1, respectively. As shown in Figure 2A, absence of Sem1 leads to an enhancement of the poor growth exhibit by each single mutant, suggesting that these factors control the same biological function. Both Sac3 and Thp1 co-purify with Ada2 (7,46). To answer whether Sem1 also co-purifies with SAGA, Ada2-TAP and UBP8-TAP were affinity purified from wild-type (wt) and sem1Δ strains. As shown in Figure 2B, Sem1 co-enriches with SAGA subunits. Nevertheless, the purification of the SAGA complex in cells lacking SEM1 did not show significant differences in the composition of the SAGA complex from wild-type cells (Figure 2C). Thus, Sem1 interacts genetically and physically with SAGA, but it is dispensable for its biochemical integrity.

In light of these genetic and physical interactions between Sem1 and SAGA, we further investigated the role of Sem1 in transcription. Thus, DNA microarray gene expression profile was generated for sem1Δ and was compared with the profiles of the mutants in the components of TREX-2 and SAGA, all of which were generated identically (39). Supplementary Figure S1A shows a hierarchical cluster diagram of the color-coded profiles according to complex composition. The dendrogram of similarities indicates that the gene expression changes observed in sem1Δ correspond most closely to those of sus1Δ. These profiles are also similar to the expression changes observed in the mutation of TREX-2 components THP1 and SAC3 (see Supplementary Figure S1B for all pairwise correlations). An unsupervised network recapitulates all these correlations and positions Sus1 as playing an intermediary role between SAGA and TREX-2 (Figure 2D). Interestingly, the sem1Δ transcriptional profile also correlates with ada2Δ to a lesser extent. Moreover, our analyses show that the expression changes of sem1Δ are enriched by genes with TATA-containing promoters with a P-value of <0.0025, these generally being SAGA-dominated genes (48,49). In summary, these results suggest that Sem1 plays a role in transcription in a similar manner to Sus1.

Sem1 is required for the induction of SAGA-regulated genes GAL1 and ARG1

Our results suggest that Sem1 could play a role in the transcription of SAGA-dependent genes. Induction of SAGA-regulated genes, such as GAL1 and ARG1, has been shown to depend on Sus1 (7,50); thus, the requirement of Sem1 for the induction of both genes was also investigated. Figure 3A shows that the induction of ARG1 by the addition of sulfometuron (SM) or GAL1 on a 25-min shift to galactose-containing medium does not reach the wild-type expression levels in cells lacking SEM1. To better follow the effect of loss of Sem1 on the GAL1 expression, GAL1 mRNA abundance was monitored during galactose induction at different time points. The absence of SEM1 reduces GAL1 induction after a shift to galactose (Figure 3B). In accordance with poor GAL1 induction, the cells lacking SEM1 displayed a 3-fold depletion of RNA pol II at the GAL1 promoter and in coding regions on the addition of galactose (Figure 3C). Thus, RNA pol II recruitment to GAL1 diminished in those cells lacking SEM1. As expected with inefficient GAL1 induction, sem1Δ mutant cells grew poorly on galactose-containing plates (Figure 3D). Interestingly, this phenotype was greatly enhanced by raising the growth temperature to 37°C (Figure 3D lower right), which correlated with a downregulation of the GAL1 mRNA expression levels in sem1Δ at that temperature (Figure 3E).

Sem1 is required for SAGA integrity at the GAL1 promoter

As shown earlier in the text, Sem1 is required for the expression of some SAGA-dependent genes. A straightforward explanation would be that SAGA components are not efficiently recruited to the GAL1 gene in sem1Δ cells. To investigate this notion, the recruitment of several SAGA subunits, which are the components of distinct modules, was monitored. Recruitment of the components of the HAT/CORE (Ada2 and Gcn5), SA_TAF (Taf9), SA_SPT (Spt8) and the DUB (Sus1) modules was analyzed. As shown in Figure 4A, the presence of HAT/CORE subunits Ada2 and Gcn5 is significantly reduced in the absence of Sem1. In addition to the poor recruitment of the HAT/CORE subunits, the presence of Taf9 (Figure 4B) and Spt8 (Figure 4C) from SA_TAF and SA_SPT, respectively, also diminished at GAL1 in sem1Δ cells. However, the recruitment of Sus1, to the same DNA region, was not significantly affected under similar conditions (Figure 4D). In summary, the ChIP results shown herein suggest that Sem1 influences SAGA integrity at the GAL1 promoter, but not Sus1 association.

Sem1 influences SAGA-dependent deubiquitylation activity

As shown earlier in the text, Sus1 was associated with GAL1 on the shift to galactose in sem1Δ, whereas other SAGA subunits are absent. However, under our experimental conditions, we were unable to demonstrate whether Ubp8 recruitment to the GAL1 promoter was affected in absence of SEM1. Thus, we ought to investigate whether SAGA-mediated DUB activity at GAL1 was compromised. To monitor the in vivo H2B ubiquitylation status at the GAL1 promoter, real-time qPCR after ChDIP was performed. Figure 5A shows that the cells lacking SEM1 displayed a 3-fold enrichment of ubiquitylated H2B at the GAL1 promoter. Levels of ubiquitylated histone H2B correlated with the H3K79 trimethylation status at the GAL1 promoter (51–54). Accordingly, H3K79 trimethylation levels abnormally increased at the GAL1 promoter on gene activation in sem1Δ cells (Supplementary Figure S2). Hence, Sem1 is required for in vivo SAGA-mediated histone H2B deubiquitylation and for H3K79 trimethylation at the GAL1 promoter.

The biochemical analyses shown in Figure 2C indicate that Sem1 is not necessary for SAGA integrity. However, at GAL1 promoter, Sem1 is required for SAGA-mediated deubiquitylation of H2B (Figure 5A). To further characterize whether the DUB module is partially inactive in sem1Δ cells, in vitro H2B deubiquitylation assays, with the SAGA complex purified via Sus1-TAP from wild-type and sem1Δ cells, were carried out. As Sgf73 is required for SAGA DUB activity, Sus1 purified via TAP in sgf73Δ was used as a negative control (20,46,55,56). Figure 5B shows that SEM1 deletion reduced in vitro H2B Sus1-dependent DUB activity by decreasing in 40% level of H2B-deUB. Moreover, the level of H2B-deUB is reduced in 60% when the complex was purified by Ubp8-TAP from sem1Δ (Figure 5C and Supplementary Figure S3). Thus, these results indicate that Sem1 enhances SAGA-mediated histone H2B deubiquitylation activity in vitro.

TREX-2 is involved in enhancing SAGA-dependent deubiquitylation activity

Sem1 affects the recruitment of different SAGA subunits and enhances SAGA-mediated histone H2B deubiquitylation. Whether this is related to its role in stabilizing TREX-2 is unknown. To test whether the mutation in other TREX-2 subunit also lead to similar phenotypes, in vitro deubiquitylation activity, GAL1 induction and SAGA recruitment were investigated in a thp1Δ strain. As shown in Figure 6A, the absence of Thp1 leads to a significant reduction of Ubp8-mediated DUB activity, similarly to that observed for Sem1 deletion (Figure 5C). Moreover, as observed for sem1Δ, the absence of THP1 impairs GAL1 induction (Figure 6B). Finally, the ChIP experiments done to monitor recruitment of different SAGA factors to the GAL1 promoter in the wild-type and thp1Δ on the shift to galactose reveal that THP1 deletion decreased the correct recruitment of distinct SAGA subunits to the GAL1 promoter (Figure 6C and D). As it happens for sac3Δ (50), the presence of Sus1 is also compromised in the thp1Δ under the same conditions (Figure 6E). In conclusion, TREX-2 subunit Thp1 is needed for a precise SAGA recruitment to GAL1. All in all, the data herein suggest that Sem1 influences the GAL1 expression by a mechanism that interferes with SAGA-DUB-mediated activity and by stabilizing the TREX-2 complex.