Gcn5 and Ubp8 dependent ubiquitylation affects glycolysis in yeast

Many of the molecular mechanisms affected by ubiquitylation are highly conserved from yeast to humans and are associated to a plethora of diseases including cancers. To elucidate the regulatory role of epigenetic factors such as the catalytic subunits of SAGA complex, KAT-Gcn5 and Ub-protease Ubp8, on ubiquitylation of non-histone proteins we have performed a comprehensive analysis of the Ub-proteome in yeast Saccharomyces cerevisiae in strains disrupted in Gcn5, Ubp8 or both respect to wild type. We found significative alteration of ubiquitylation in proteins belonging to different functional categories with a recurrence of identical proteins in absence of Gcn5 or Ubp8 indicating shared targets and their interlaced function. Among the processes involved we noteworthy identified all major enzymes engaged in energy metabolism and glycolysis such as PFK1, PFK2 and others showing increased ubiquitylation respect to WT. We showed that the higher degree of ubiquitylation found is at post-translational level and does not depend on transcription. Noteworthy, we found in vivo severe defects of growth in poor sugar medium and inability to adaptive switch from fermentative to respiratory growth in strains lacking Gcn5 and Ubp8. Our findings data provide a novel, direct link, between metabolism and epigenetic control with a novel role of DUB-Ubp8 and KAT-Gcn5 on the ubiquitylation marking all the main glycolytic enzymes required for an effective execution of the glycolytic flux. Collectively our experimental results and the proposed model can lead to future research and innovative strategies that by targeting epigenetic modulators might be able to lower sugar utilization also in human cells. Author Summary Molecular mechanisms dissected in simple yeast might be translated to similar circuitries in human cells for new discoveries in human diseases including cancer. Ubiquitylation of proteins is an evolutionary conserved mechanism required for many biological processes. Different post-translational modifications (PTMs) such as ubiquitylation, acetylation, methylation etc. are reciprocally regulated for deposition or removal. Epigenetic factors writing the PTMs code are often components of multiproteic complexes such as SAGA complex that holds the K-acetyltransferase (KAT) Gcn5 and the Ubiquitin-protease (DUB) Ubp8 highly conserved in Evolution. Cells respond to environment and nutrients by changing metabolism and group of enzymes involved in specific pathways are often coregulated by the deposition of selected PTMs. This study analyses the composition and quantitation of Ub-proteins differentially modified in absence of KAT-Gcn5 and DUB-Ubp8 in yeast. Interstingly, we highlighted the role of Gcn5/Ubp8 dependent ubiquitylation in marking major glycolytic enzymes necessary for glucose utilization. Our study suggests a novel regulatory pathway and, considering that lowering glycolysis is a promising strategy to target tumor metabolism, we propose this study as an interesting perspective to lower enhanced glycolysis in tumors.

with a recurrence of identical proteins in absence of Gcn5 or Ubp8 indicating shared targets and their interlaced function. Among the processes involved we noteworthy identified all major enzymes engaged in energy metabolism and glycolysis such as PFK1, PFK2 and others showing increased ubiquitylation respect to WT. We showed that the higher degree of ubiquitylation found is at post-translational level and does not depend on transcription.
Noteworthy, we found in vivo severe defects of growth in poor sugar medium and inability to adaptive switch from fermentative to respiratory growth in strains lacking Gcn5 and Ubp8.
Our findings data provide a novel, direct link, between metabolism and epigenetic control with a novel role of DUB-Ubp8 and KAT-Gcn5 on the ubiquitylation marking all the main glycolytic enzymes required for an effective execution of the glycolytic flux. Collectively our experimental results and the proposed model can lead to future research and innovative strategies that by targeting epigenetic modulators might be able to lower sugar utilization also in human cells.

Author Summary
Molecular mechanisms dissected in simple yeast might be translated to similar circuitries in human cells for new discoveries in human diseases including cancer. Ubiquitylation of proteins is an evolutionary conserved mechanism required for many biological processes.
Different post-translational modifications (PTMs) such as ubiquitylation, acetylation, methylation etc. are reciprocally regulated for deposition or removal. Epigenetic factors writing the PTMs code are often components of multiproteic complexes such as SAGA complex that holds the K-acetyltransferase (KAT) Gcn5 and the Ubiquitin-protease (DUB) Ubp8 highly conserved in Evolution. Cells respond to environment and nutrients by changing metabolism and group of enzymes involved in specific pathways are often coregulated by the deposition of selected PTMs. This study analyses the composition and quantitation of Ub-proteins differentially modified in absence of KAT-Gcn5 and DUB-Ubp8 in yeast. Interstingly, we highlighted the role of Gcn5/Ubp8 dependent ubiquitylation in marking major glycolytic enzymes necessary for glucose utilization. Our study suggests a novel regulatory pathway and, considering that lowering glycolysis is a promising strategy to target tumor metabolism, we propose this study as an interesting perspective to lower enhanced glycolysis in tumors.

Introduction
Conservation in Evolution allows translation of discoveries obtained in simple models useful for the interpretation of more complex issues in human cells. Accordingly, the simple budding yeast is one of the best models to study molecular mechanisms regulating complex networks and circuitries [1]. Proteins are differentially regulated by a plethora of posttranslational modifications (PTMs) often involved in a reciprocal cross talk and regulation. Lysine acetylation, for example, prevents further poly-ubiquitylation on the same residue with a direct impact on several functions including protein turn over and proteasomal degradation [2]. The big multiproteic SAGA complex (Spt-Ada-Gcn5-acetyltransferase) is an epigenetic key regulator of acetylation and establishes the transcriptional competence of gene promoters through the opening of chromatin to the transcriptional machinery [3]. Its relevance is not merely confined to histones modifications but extend far behind to the deposition of PTM marks on non-histone proteins. It is composed of 21 widely conserved proteins grouped into functional submodules, it harbours K-acetylatransferase KAT2-Gcn5 mediating acetylation and Deubiquitylase (DUB) Ub-protease Ubp8 mediating the deubiquitylation [4,5]. The coexistence in SAGA complex of acetylation and deubiquitylation modules sustains their possible interdependency and interlaced functions. Gcn5 and Ubp8 indeed showed functional interaction in fermentative or respiratory growth conditions [6] and at high and low temperatures previously reported to be correlated with glycolysis [7]. In the present study we investigated polyubiquitylated proteins expressed in absence of KAT-Gcn5, DUB-Ubp8 or both in comparison with wild type strains. We approached this study by expressing His6-Ub in the different yeast strains performing selective biochemical purification of tagged Ub-proteins. After elution we analyzed the composition of Ub-proteins obtained with a gel-free and label-free proteomic approach based on the coupling of micro-liquid chromatography and tandem mass spectrometry (µLC-MS/MS). Certainly, proteomics is proving its potential as tool of choice to reach a deepened characterization of proteinprotein interactions and PTMs that regulate crucial biological processes. Using a multidimensional chromatographic system coupled to mass spectrometry it was investigated, for example, the function of TFIID in S.cerevisiae by identifying new connections between its components and novel subunits of SAGA complex obtaining the most complete proteomic interaction map of an eukaryotic transcription machinery [8].
Recently, Wilson M.A. et al. [9] demonstrated with a proteomic approach that SAGA complex also deubiquitylates SNF1 altering its kinase activity confirming that the acetylation of non-histone proteins plays a direct role in the regulation of protein functions. This evidence is in agreement with results obtained in a study on Gcn5 and Esa1 mutants [10] with SILAC and found their coordinated function in the regulation of cell growth and division.
It was reported that, PTMs regulate central carbon metabolism, glycolysis and fermentation where a high number of key enzymes engaged in sugar utilization and energy supply were marked by phosphorylation, acetylation and ubiquitylation indicating a direct involvement of PTMs in the regulation of their catalytic activity [11]. In addition, we have recently demonstrated that Gcn5 and Ubp8 beside their role as catalytic subunits of SAGA complex, are also involved in the control of the respiratory metabolism and mitochondrial functions [6,12]. Following this line of reasoning and starting from a shotgun proteomic profiles and in vivo phenotypic assays in strains missing Gcn5 and Ubp8 or both we wanted to deeply characterize the composition of differentially ubiquitylated proteins in order to clarify their role and investigate possible correlation with the regulation of sugar metabolism. Our experimental data provided significative results thus demonstrating an altered ubiquitylation on main glycolytic enzymes in absence of Gcn5 and Ubp8. We believe that our results shed light on a novel role of SAGA complex in the ubiquitylation of metabolic enzymes opening to new perspectives on the relevance of ubiquitin in regulating glucose metabolism also in higher cells. Therefore, the modulation of PTMs on key enzymes may lead to a new strategy to slow or block sugar metabolism that may counteract the rapid and energy demanding proliferation of cancer cells.

Genetic and functional interaction between HAT-Gcn5 and DUB-Ubp8 in S.cerevisiae.
Based on the coexistence in the multiproteic SAGA complex of the KAT and DUB modules and the deubiquitylation activity of Ubp8 on the centromeric histone variant Cse4 previously reported [12] we wanted to analyse whether the global process of protein ubiquitylation might be dependent on Gcn5 or Ubp8. In S.cerevisiae, the functional interaction between genes is tested at genetic level by comparison of the growth of single disrupted strains respect to the strain carrying the double genes disruption in order to analyze the resulting synthetic phenotype and the fitness of the strains grown in selective conditions (Fig. 1A).
We previously demonstrated the genetic interaction between Gcn5 and Ubp8 in respiratory growth conditions [6,12]. Here we have assayed S.cerevisiae gcn5∆, ubp8∆ and ubp8∆/gcn5∆ strains and confirmed their genetic interaction by growing at low (12°C), physiological (28°C) and high (35°C) temperatures (Fig. 1B). Growth spot assay suggested that there is a severe growth defect at 12°C and at 35°C with a remarkable synthetic defective phenotype in the growth of the double disrupted ubp8∆/gcn5∆ strain. Based on these genetic evidences we decided to investigate if loss of Ubp8 and Gcn5 or both might, possibly, affect the ubiquitylation process of non-histone proteins in an interdependent way.
The 6His-Ub-proteins were expressed in WT, ubp8∆, gcn5∆ and ubp8∆/gcn5∆ strains according to procedure described in material and methods ( Fig. 2A), purified on Ni-column [13] and analyzed on western blot. Fig. 2B shows the selective purification of the obtained His6Ub-proteins, evidenced with anti-His6 antibody (Fig. 2B). Unpurified, total protein extracts, indicated as input, were hybridized with anti-Ada2 antibody as an internal control indicating the expression of a constitutively expressed protein in our cellular extracts. The quality of the Ub-protein obtained prompted us to go further and perform a proteomic characterization after trypsin digestion of 6His-Ub-proteins retained by means of a shotgun label-free platform. In particular, our proteomic approach, based on a µLC-MS/MS, allowed the identification of 447 distinct proteins, whose distribution in the four analyzed strains is represented by a Venn diagram shown in Fig. 2C. A total of 16 µLC-MS/MS runs were acquired, evaluating two preparations for the four experimental conditions (biological replicates) and analyzing them in duplicate (technical replicates). The complete list of proteins identified for each group is reported in Table S1. Considering the strains disrupted in KAT-Gcn5, in the Ub-protease Ubp8 or in both (gcn5 ubp8 and ubp8/gcn5) compared to WT, we observed that the majority of proteins identified in ubp8∆ and gcn5∆ strains were also found in the double ubp8∆/gcn5∆ strain. In fact 193 proteins, corresponding to the 43,2% of total, were shared among the four strains, suggesting that these proteins represent common targets of both, Gcn5 and Ubp8. Moreover, 48 proteins have been identified in common between WT and ubp8 and 18 between WT and gcn5, suggesting that Ubp8 is involved in the regulation of more proteins respect to Gcn5 (10,7% vs 4%). This observation was expected and based on the Ub-protease activity of Ubp8, whereas we expect that Gcn5 acts indirectly possibly through the deposition of a counteracting acetyl group on specific lysines. Based on the general results obtained and recapitulated in the Venn diagram ( Fig. 2C), we decided to investigate in more depth the differences in expression of ubiquitylated proteins in the examined strains by performing label-free differential analyses.

Label-free differential analysis
Using MAProMa software [14] all the sixteen protein lists were aligned and for each strain a unique list has been created averaging the spectral count values (SpC*) of the identified proteins in order to estimate their relative abundance in the pairwise comparisons with gcn5, ubp8 and ubp8/gcn5 strains respect to the WT strains. To this end, the two algorithms of MAProMa software, DAve and DCI, were applied, representing the ratio and the confidence in differential expression, respectively, for each protein between two samples. Using stringent filters described in M&M section (see pag. XX) to maximize the confidence of identification, a total of 107 proteins were found differentially expressed and reported in Table 1 with selected details and in Suppl. Table S1 in extended form. In particular, 6 and 46 proteins resulted respectively down-and up-represented in ubp8 compared to WT, while, 48 proteins were more abundant in gcn5 compared to WT and 8 were down-represented in gcn5 strain. Finally, 12 and 64 proteins resulted downand up-represented, respectively in ubp8/gcn5 strain compared to WT. In addition, known genetic (G) or physical (P) interactions with, respectively, Ubp8 or Gcn5 are reported, and, finally, the corresponding name of the human hortholog gene. The trend of the abundance of proteins among the pairwise comparisons of the three mutant strains respect to WT is shown through a color code corresponding to the individual DAve values assigned: for DAve values ranging from +0.40 to +2.00 color gradations are used from light red to dark red and for those ranging from −2.00 to −0.40 chromatic shades from dark blue to light blue. Moreover, in order to give a more comprehensive view of the pathways and of the proteins involved with KAT-Gcn5 and DUB-Ubp8 interactions, in Table 1 a color code is assigned also to those proteins whose DAve and DCI values do not pass the filters in all the considered comparisons. So, proteins with DAve values ranging from -0.4 to +0.4 appear with light red and light blue coloring respectively, while the unidentified or unchanged ones are shown with white code. Therefore, observing the table and the trends of the differentially expressed proteins among the three comparisons, it is immediately evident that some proteins seems to be more ubiquitylated in absence of DUB-Ubp8 others of KAT-Gcn5 indicating the major regulatory role of Ubp8 and Gcn5 on the specific protein. For example, ASN2, LEU1, PYC2, TDH2, GSY2, PGM2 and TY1B-PR1 are proteins that equally change in the gcn5 and in the double mutant strain, these evidences suggest that they are closely related to Gcn5 as much more influenced by the mutation of this gene. In the same way, YEF3, CDC19, CYT1, HSC82 and YHR097C equally change in the ubp8 and in the double mutant strain, thus demonstrating that they are closely related to Ubp8 as much more influenced by the mutation of this gene. Another interesting aspect concerns those proteins (such as, EFT1, PGI1, GPD1, SOD1, AMD1 and URA2) that show a marked increase or reduction in their levels if we consider the ubp8/gcn5 strain compared to the single ones, demonstrating in this case the synergistic effect exerted by the double deletion of Ubp8 and Gcn5. Finally, in Table 1 we can also observe other proteins (ILV6, PFK1, PFK2, SCH9, HNT1, GCV2) that show the same differential trend both in the single and in the double mutant strains. Noteworthy the physical (P) or genetic (G) interactions of Ub-proteins with either Gcn5 and Ubp8 supports these results, indicating a direct effect of Gcn5 and Ubp8 in their modification. Among these, we found proteins that can be grouped on the basis of their functional role, such as amino acids biosynthesis, glycolysis, fermentation, oxidative phosphorylation and energy metabolism. With the aim to better visualize the pathways and the biological relevant interactions in which these proteins are involved in, the STRING database [15] was queried to build a network of both known and predicted protein-protein interactions. Differentially levels of proteins were grouped in ten sub networks based on their molecular function and represented by nodes and grey edges indicating protein-protein interactions. Applying the same chromatic scale of Table 1, Figure 3 show the three networks involving the differentially expressed protein behaviors towards the pairwise comparisons of gcn5, ubp8 and ubp8/gcn5 strains against WT condition. Differentially expressed proteins resulted from the MAProMa comparison of ubp8∆, gcn5∆ and ubp8∆-gcn5∆ strains versus WT condition. In particular, each protein (identified in figure through its Gene Name) is marked by a color code which it is defined by the DAve value obtained in the three examined comparisons. The color is assigned according to a chromatic scale (reported in figure) which represents the confidence ranges of DAve values adopted (from -2.00 to 0 a gradient from blue to white and from +2.00 to 0 a gradient from red to interactions with Ubp8 and/or Gcn5 and the name of the human hortholog. In red the yeast genes corresponding to human horthologs related to human pathologies. The complete list of the reported proteins was extracted from the differential lists in Table S2.

Major glycolytic enzymes are differentially ubiquitylated in absence of Gcn5 or Ubp8
The fact that several altered proteins found in our screening are linked to glycolysis, oxidative stress and energy metabolism (see Fig.3) could support the hypothesis that glycolysis is impaired in cells lacking Gcn5 and Ubp8. This observation, previously reported by Tripodi et al. [11], would strengthen our proteomic data and sustain a regulatory role of ubiquitylation on fundamental glycolytic enzymes. We report in Fig. 4A the scheme of the major steps composing the glycolytic flux for the utilization of glucose bringing to production of pyruvate from glucose-6-P. Major enzymes involved in these metabolic reactions were identified in our screening and Fig. 4A shows the proteins colored according to the color code previously described. This finding sustain a strong effect on the ubiquitylation level on the most important enzymes involved in glycolysis suggesting an impairment in the glycolytic flux progression. In Fig. 4B is reported the detailed color code of each identified enzyme found in strains disrupted in Gcn5, Ubp8 or both compared with the wild type. The panel provides the overall picture demonstrating how strong is the differential level of ubiquitylation on these proteins. We therefore asked whether the transcriptional expression was also affected by Gcn5 or Ubp8 disruption, and we made the analysis of the PFK1 and PFK2 mRNA expression, heavily ubiquitylated in all the strains and PYC1 showing only a slight difference between strains disrupted in Gcn5 or Ubp8. Fig. 4C shows diagrammed results obtained with RT qPCR on mRNA expression demonstrating no effects of either Gcn5 or Ubp8 disruptions on mRNA expression of the analyzed genes. Collectively, these results exclude any up-regulation at transcriptional level of these genes confirming that the higher degree of ubiquitylation found is at post-translational level and does not depend on transcription.

Glycolysis is impaired in absence of Gcn5 and Ubp8
In order to assay their efficiency in sugar utilization, we then decided to grow WT, gcn5, ubp8 and the double ubp8/gcn5 strains in medium with high (2%) and low (0,1%) glucose in order to assay their efficiency in sugar utilization. According to our model, Fig. 5A shows a severe defect of growth in absence of Gcn5 and Ubp8 in low glucose that is not recovered by addition of downstream products such as lactate and ethanol. We repeated this analysis following prolonged growth in rich medium liquid cultures in the presence of 2% and 0,2% glucose. A strong impairment of growth in absence of Gcn5 and Ubp8 was observed after 24 hours in low sugar, while in the wild type there was only minor effects of sugar concentration on growth. Disrupted strains were unable to duplicate after 24 hours and growth was blocked, indicating inability to switch to respiratory conditions (Fig. 5B). To confirm these data, we used the Alcohol-dehydrogenase activities, Adh1 and Adh2, as respiro-fermentative glycolytic flux markers. In fact, Adh1, which is primarily involved in ethanol production, is expressed in fermentative, high glucose, conditions [16]. On the other side, Adh2 is activated when glucose is exhausted and the cell switches to respiratory metabolism by growing mainly on the accumulated ethanol. The transition from one condition to the other can be followed by the native in-gel appearance of the glucoserepressible ADH2 gene product with an Adh-specific assay, previously used to dissect the ADH genetic system [17] [18] [19]. To this end, protein extracts from 2% and 0,2% glucosecontaining cultures of WT, ubp8, gcn5 and ubp8/gcn5 strains grown for 30, 60 and 90 hours were fractioned on native PAGE and stained for Adh. It must be underlined that the use of 0,2% glucose was a necessary condition to avoid the nearly complete impairment of growth of gcn5 and ubp8 strains in more stringent 0,1% glucose-containing medium. The Adh pattern of the wild type on 2% and 0,2% glucose can easily describe the expected growth phases (Fig. 5C). obtained from the four examined strains and then we achieved quantitative information on specific proteins belonging to cellular metabolic processes subjected to fast changes in response to genetic or environmental stimuli. In fact, as stated above, unlike transcriptomics, our proteomic approach was able to extract the trend of relative abundance of differentially expressed proteins among the three mutated strains compared to WT strain.
In addition, it was possible to construct a complete proteomic interaction map to better visualize the major pathways in which these proteins resulted involved. In the first instance, the results obtained clearly indicated a central role for Ubp8 and Gcn5, affecting specific proteins that contribute to the control of cell growth, stress response and above all energetic metabolism. The comparison of Ub-proteins in WT, gcn5∆, ubp8∆ and ubp8/gcn5 strains identified proteins whose ubiquitylation vary significatively thus indicating their bona fide targets. Moreover, most of the Ub-proteins found as differentially expressed in the examined comparisons, although not always with the same levels of abundance, resulted shared and up-represented in absence of Ubp8, Gcn5 and in the double disrupted strain. These evidences can be interpreted as due to modification of lysines that can alternatively be acetylated or ubiquitylated [11], or, even, to a lower expression of Ubp8 in gcn5 strain [6].
As also reported by Wilson et al. [9], our data confirmed that SAGA modulates the PTMs of SNF1, involved in the inactivation of enzymes of fatty acid biosynthesis and glycogen storage and it is a positive regulator of autophagy [21], accordingly and found a lower ubiquitylation of this protein in the ubp8 strain. After grouping these Ub-proteins respect to their functional category we observed the presence of major enzymes involved in glycolysis.
Noteworthy, the same proteins were described to be ubiquitylated/acetylated [11].
Remarkably, Pfk1 and Pfk2 key nodes for the execution of the glycolytic flux [22] remarkably, there was no upregulation at transcriptional level of these genes confirming that the higher degree of ubiquitylation found is at post-translational level. It can not be ruled out that an additional role of ubiquitylation might be also involved in a previously proposed allosteric regulation of the enzymatic activity of Pfk1 and Pfk2 [23]. Other genes determining the final utilization of glucose, were identified in our screening (Tab1 and Fig. 3). In the presented model we indicate alternative metabolic redox-balancing routes undertaken in absence of Gcn5 and Ubp8. We report a biochemical analysis of strains growth in high and low sugar where Adh1 and Adh2 activities were used as metabolic marker of fermentation and respiration [24]. Growth was severely impaired in 0,2% glucose accordingly, an extremely low fermentative potential was correlated to the exclusive presence of Adh1 (Fig.   5C) in ubp8 strain demonstrating its full inability to switch to respiratory conditions as reported in a previous reports [6]. On the contrary, gcn5 strain displayed weak amount of Adh2 in fermentative conditions (2% glucose) suggesting the rerouting of the glycolytic flux towards a respiro-fermentative metabolism (Fig. 6). We believe that this is probably due to a partial reoxidation of Adh1-dependent accumulation of ethanol by small amount of Adh2, sufficient to transfer the cytoplasmic NADH redox excess to the mitochondrial compartment outer mitochondrial membrane transdehydrogenases (Nde1 and Nde2) involved in glycolytic fermentation to ethanol (Fig. 6) [25]. In this respect we suggest that the differential capacity of external and internal NADH dehydrogenases in the two deleted strains, interpreted in terms of redox metabolism, plays crucial roles in triggering of fermentation [20]. In low 0,2% glucose there is a complete lack of Adh2 that is usually regulated by Adr1 repressed in high glucose [26]. In addition to Pfk1 and Pfk2, showing highest levels ubiquitylation in absence of Gcn5 and Ubp8, other important enzymes such as Tdh2, Pyc1 and Cdc19 at the crossroads of glycolysis, gluconeogenesis and pentose phosphate pathways, showed altered ubiquitylation in both deleted strains (Fig. 4A, B). On the basis of our collected results, we can speculate, that an altered ubiquitylation affects the function of many This is not surprising since lysine acetylation often counteract further ubiquitylation suggesting that there is a reciprocal interlaced role of these factors [2]. We think that the presented results may open the way for a novel framework of research with interesting implications for the study of cancer cells where increased glucose utilization and altered glycolytic flux occur independently to oxygen availability [27] [28]. In case of proinflammatory stimuli, for example, the induction of a metabolic switch leading to upregulation of aerobic glycolysis [29] lead to an enhanced utilization of glucose which sustains the high rate of cell proliferation and invasiveness [30]. The role of Gcn5 and Ubp8 affecting protein ubiquitylation in yeast is a novel finding shedding light on the importance of the epigenetic post-translational regulation not only for the regulation of nucleosomes accessibility but also at metabolic level for sugar utilization. In sum, we show a direct requirement of the SAGA acetyltransferase Gcn5 and Ub-protease Ubp8 in the regulation of this process strengthening previously data indicating the presence of acetylated and ubiquitylated version among glycolytic enzymes [11]. As a possible translational application of our results we recall the role of the "cancer signature gene" Usp22, hortholog of Ubp8, in aggressive tumors such as kidney and glioma [31]. We might therefore envisage novel strategies that may alter energy supply and lower glycolysis by targeting SAGA components as novel approaches to slow proliferation and invasion of cancer cells.

Yeast strains and growth
Saccharomyces cerevisiae strains derive from the isogenic WT W303 and are listed in Table strains. Gene disruptions of GCN5 and UBP8 or both was carried out with integrative marker cassette and controlled for in situ integration by colony PCR. His6Ub was expressed from pDJ421 plasmid. Growth at 28°C, in YP medium (1% yeast extract, 2% bactopeptone, 1% agar) containing 2%, 0.2%, and 0,1% glucose, 3% glycerol, plus 1% EtOH or 1% lactate. Growth assay: 1 OD 600 of exponentially growing cells was serially diluted (1/5) and spotted.

Expression, biochemical purification of his6-Ub proteins
Cells transformed with pDJ421 encoding 6His-ubiquitin under the CUP1 inducible promoter P CUP1 , [32] were grown on selective media and activated overnight with 0.1 mM CuSO4.

SDS-PAGE, western blot and immunohistochemistry
Yeast cultures grown at 28 °C semi-complete medium (0,67% yeast nitrogen base, 2% glucose, 0.1% drop out mix). For expression of plasmid 0.1mM of CuSO 4 was added to culture. Cells were collected at exponential phase (0.8 OD600/ml). Eluates and lysate were loaded on 7.5% SDS-PAGE run and blotted on nitrocellulose membranes (Amersham).

Real time qRT-PCR
Exponentially growing cells (WT, gcn5Δ and ubp8Δ strains) were collected after overnight growth in YP-glu2%. Total RNA extraction was performed using phenol method and retro- based on q-values, considering maximum deltaCN of 0.05 [33]. Only peptides with high confidence, minimum peptide length of six amino acids, and rank 1 were considered.
Protein grouping and strict parsimony principle were applied. The output data obtained from SEQUEST software, i.e., Spectral counts (total number of spectra identified for each protein), were treated with MAProMA (Multidimensional Algorithm Protein Map), an in-house algorithm for comparison of protein lists, evaluation of relative abundances, and plotting of virtual 2D maps [14]. . It should be noted that we were unable to assign the cellular function for some proteins which were therefore classified as "unknown" and additional proteins without a complete characterization at the moment of data analysis were classified as "unclear" or "other". Label-free differential analysis. The sixteen protein lists obtained from the SEQUEST algorithm were aligned and compared by means of the average spectral counts (aSpCs) corresponding to the average of all the spectra identified for a protein and, consequently, to its relative abundance, in each analyzed condition (WT,   (yellow) gcn5∆ (green) and ubp8-gcn5∆ (red). Area of intersection contain proteins common to different strains.  Supplementary Table S2 for the complete reference), represented as a node and grouped to others belonging to the same pathway. Only experimentally and computationally predicted protein-protein interactions are considered and indicated as grey edges. The color code of distinct nodes represents the DAve value and the relevant chromatic scale (reported in figure) ranges from -2.00 to 0 (dark blue to white) and from 0 to +2.00 (white to dark red).
Proteins with DAve ≥ l0.4l and a DCI ≥ l6l pass the filters and could be considered differentially expressed in the considered comparison.   Supporting Informations Legends S1 Table. Complete list of the Ub-proteins detected in WT, ubp8∆, gcn5∆ and ubp8∆-gcn5∆ S.cerevisiae strains.
In this table for each protein is reported: Uniprot Accession, Reference, Gene name, pI, MW, Spectral Counts (SpCs), Score and Frequency. Frequency indicates how many times a given protein has been identified in the four replicates under each examined condition.
The asterisk next to SpC and Score indicates that the average values are given for each protein of the same condition. Methods section. It should be noted that proteins were primarly grouped accordingly to their Molecular function and secondarily by their Gene name.