Interplay between Sumoylation and Phosphorylation for Protection against α-Synuclein Inclusions*

Background: Phosphorylation and sumoylation are post-translational modifications of the Parkinson disease protein α-synuclein. Results: α-Synuclein inclusion clearance is impaired in yeast when sumoylation is inhibited; phosphorylation of α-synuclein can compensate SUMO impairment. Conclusion: Sumoylation stimulates autophagy clearance of α-synuclein inclusions, whereas phosphorylation promotes autophagy and proteasome degradation. Significance: A complex molecular post-translational cross-talk is required in yeast to clear toxic inclusions. Parkinson disease is associated with the progressive loss of dopaminergic neurons from the substantia nigra. The pathological hallmark of the disease is the accumulation of intracytoplasmic inclusions known as Lewy bodies that consist mainly of post-translationally modified forms of α-synuclein. Whereas phosphorylation is one of the major modifications of α-synuclein in Lewy bodies, sumoylation has recently been described. The interplay between α-synuclein phosphorylation and sumoylation is poorly understood. Here, we examined the interplay between these modifications as well as their impact on cell growth and inclusion formation in yeast. We found that α-synuclein is sumoylated in vivo at the same sites in yeast as in human cells. Impaired sumoylation resulted in reduced yeast growth combined with an increased number of cells with inclusions, suggesting that this modification plays a protective role. In addition, inhibition of sumoylation prevented autophagy-mediated aggregate clearance. A defect in α-synuclein sumoylation could be suppressed by serine 129 phosphorylation by the human G protein-coupled receptor kinase 5 (GRK5) in yeast. Phosphorylation reduced foci formation, alleviated yeast growth inhibition, and partially rescued autophagic α-synuclein degradation along with the promotion of proteasomal degradation, resulting in aggregate clearance in the absence of a small ubiquitin-like modifier. These findings suggest a complex interplay between sumoylation and phosphorylation in α-synuclein aggregate clearance, which may open new horizons for the development of therapeutic strategies for Parkinson disease.

Parkinson disease is associated with the progressive loss of dopaminergic neurons from the substantia nigra. The pathological hallmark of the disease is the accumulation of intracytoplasmic inclusions known as Lewy bodies that consist mainly of post-translationally modified forms of ␣-synuclein. Whereas phosphorylation is one of the major modifications of ␣-synuclein in Lewy bodies, sumoylation has recently been described. The interplay between ␣-synuclein phosphorylation and sumoylation is poorly understood. Here, we examined the interplay between these modifications as well as their impact on cell growth and inclusion formation in yeast. We found that ␣-synuclein is sumoylated in vivo at the same sites in yeast as in human cells. Impaired sumoylation resulted in reduced yeast growth combined with an increased number of cells with inclusions, suggesting that this modification plays a protective role. In addition, inhibition of sumoylation prevented autophagymediated aggregate clearance. A defect in ␣-synuclein sumoylation could be suppressed by serine 129 phosphorylation by the human G protein-coupled receptor kinase 5 (GRK5) in yeast. Phosphorylation reduced foci formation, alleviated yeast growth inhibition, and partially rescued autophagic ␣-synuclein degradation along with the promotion of proteasomal degradation, resulting in aggregate clearance in the absence of a small ubiquitin-like modifier. These findings suggest a complex interplay between sumoylation and phosphorylation in ␣-synuclein aggregate clearance, which may open new horizons for the development of therapeutic strategies for Parkinson disease.
Parkinson disease (PD) 3 is the second most common neurodegenerative disorder after Alzheimer disease. Pathologically, it is characterized by loss of dopaminergic neurons in the substantia nigra pars compacta of the brain and the accumulation of cytoplasmic inclusions termed Lewy bodies (1,2). Lewy bodies are composed of different proteins such as ␣-synuclein, ubiquitin, Synphilin-1, or cytoskeletal proteins (3). The small neuronal protein ␣-synuclein (␣Syn) consists of 140 amino acids and represents the major component of Lewy bodies (4). In addition, mutations (5-7) and multiplications (8) of the SNCA gene, coding for ␣Syn, cause familial forms of PD, further supporting the involvement of ␣Syn in pathogenesis. However, the precise molecular mechanisms underlying ␣Syn toxicity are still unclear. Several studies reported that ␣Syn is subjected to various post-translational modifications that can alter ␣Syn inclusion formation and cytotoxicity (9). These include sumoylation, phosphorylation, ubiquitination (10 -12), or nitration (13,14).
It has been shown that sumoylation negatively regulates ␣Syn aggregation by promoting its solubility (15). Besides ␣Syn, there are additional examples of proteins involved in neurodegenerative diseases that are SUMO targets (16,17). The predominant ␣Syn phosphorylation site (Ͼ90%) is serine 129 (Ser-129) in Lewy bodies (18,19). Several kinases such as G protein-coupled receptor kinases or Polo-like kinases 1-3 and casein kinases 1 and 2 can phosphorylate ␣Syn on Ser-129 in human cells (18 -24). Phosphorylation of ␣Syn by GRK5 plays a crucial role in the pathogenesis of PD (25). PLK2 is the most efficient Polo-like kinase phosphorylating ␣Syn on Ser-129 (26 -28). The role of ␣Syn phosphorylation under physiological conditions and in inclusion formation and pathogenesis remains controversial. In Alzheimer disease, increased Tau phosphorylation can stimulate its sumoylation (29). There is also additional evidence indicating that the cross-talk between phosphorylation and sumoylation can affect substrates in different ways (30), suggesting this might also modulate ␣Syn function, distribution, and/or aggregation.
The molecular mechanisms involved in the clearance of ␣Syn aggregates is a central question for elucidating the ␣Synrelated toxicity. Soluble ␣Syn can be targeted to the 26 S proteasome for degradation (31)(32)(33)(34) or can be degraded by the autophagy-lysosomal pathway (33)(34)(35)(36). The budding yeast Saccharomyces cerevisiae has been extensively used as a powerful system to study the basic molecular mechanisms involved in ␣Syn-mediated cytotoxicity (37)(38)(39)(40). We showed that aggregate clearance of ␣Syn depends mainly on the autophagy pathway (38). Here, we addressed the question of whether the crosstalk between specific post-translational modifications of ␣Syn modulates the processing of inclusions through degradation by autophagy or the proteasome. For the first time, we demonstrate an interplay between ␣Syn sumoylation and phosphorylation to control protein turnover. ␣Syn is sumoylated in yeast cells at the same site as in human cells and can be efficiently phosphorylated on Ser-129 by the heterologously expressed human G protein-coupled receptor kinase 5 (GRK5). Interestingly, we found that sumoylation exhibits a protective role against ␣Syn toxicity and inclusion formation, and likewise, phosphorylation alleviates ␣Syn-mediated toxicity in SUMOdeficient cells by partially rescuing autophagic aggregate clear-ance and promoting proteasome-mediated degradation of ␣Syn. Altogether, our findings support that a deeper understanding of the interplay between different post-translational modifications in ␣Syn might open novel opportunities for therapeutic intervention in PD and other synucleinopathies.

EXPERIMENTAL PROCEDURES
Yeast Strains, Plasmids, Transformation, and Growth Conditions-Plasmids and S. cerevisiae strains are listed in Tables 1 and 2. Wild-type (WT) ␣Syn encoding the cDNA sequence (referred to as SNCA) or the A30P mutant sequence was cloned into the integrative pRS306 and pRS304 vectors or into the yeast pRS426 overexpression vector (41) preceded by the GAL1 promoter and followed by CYC1 terminator. The K96R/K102R mutant constructs and the S129A mutant were generated by site-directed mutagenesis using Stratagene QuikChange sitedirected mutagenesis kit (Agilent Technologies). Plasmids pME3945 and pME3597 were used as templates for generation of the desired amino acid substitutions. Human kinases GRK5 and PLK2 were cloned into the SmaI restriction site of pME2792 yeast vector proceeded by the GPD1 and GAL1 promoter, respectively. All constructs were analyzed by sequencing. For microscopy analysis, all ␣Syn variants were tagged at the C terminus with GFP via the KLID linker (38).
S. cerevisiae strains W303-1A, smt3 ts , and ulp1 ts were used for transformations performed by standard lithium acetate protocol (42). Transformations into the temperature-sensitive smt3 ts and ulp1 ts strains were performed at 25°C. All strains  were grown in synthetic complete medium (SC) (43) lacking the nutrient amino acid (uracil, histidine, or tryptophan) corresponding to the marker, and supplemented with 2% raffinose or 2% galactose. ␣Syn expression was induced by shifting yeast cells cultivated overnight in raffinose to galactose medium (A 600 ϭ 0.1). Spotting Assay-For growth test on solid medium, yeast cells were pre-grown in minimal medium containing 2% raffinose lacking the corresponding marker to mid-log phase. Cells were normalized to equal densities, serially diluted 10-fold starting with an A 600 of 0.1, and spotted on SC plates containing either 2% glucose or 2% galactose and lacking the corresponding marker. smt3 ts mutant cells were incubated at permissive temperature (25°C) and restrictive temperature (30°C). After 3 days of incubation, the plates were photographed.
Fluorescence Microscopy and Quantifications-Wild-type (W303-1A) yeast cells harboring ␣Syn were grown in SC selective medium containing 2% raffinose at 30°C and smt3 ts mutant cells at 25°C overnight and transferred to 2% galactose containing medium for induction of ␣Syn expression for 6 h. Smt3 ts mutant cells were induced at 25 and 30°C. Fluorescent images were obtained with Zeiss Observer Z1 microscope equipped with CSU-X1 A1 confocal scanner unit (YOKOGAWA), QuantEM: 512SC (Photometrics) digital camera, and Slide-Book 5.0 software package (Intelligent Imaging Innovations). For quantification of aggregation, at least 300 cells were counted per strain and per experiment. The number of cells presenting inclusions was referred to the total number of cells counted. The values are the mean of at least three independent experiments.
Immunoblotting-Wild-type (W303-1A) yeast cells harboring ␣Syn were pre-grown at 30°C in SC selective medium containing 2% raffinose. Cells were transferred to SC medium containing 2% galactose at A 600 ϭ 0.1 to induce the GAL1 promoter for 5 h. Smt3 ts cells harboring ␣Syn were preincubated at 25°C and later transferred to either 25 or 30°C. Total protein extracts were prepared, and the protein concentrations were determined with a Bradford assay. 10 g of each protein were subjected to 12% SDS-PAGE and transferred to a nitrocellulose membrane. Membranes were probed with mouse ␣Syn monoclonal antibody (AnaSpec, Fremont, CA), rabbit ␣Syn polyclonal antibody (Santa Cruz Biotechnology), or SUMO rabbit antibody (Rockland Immunochemicals Inc.). Rabbit cdc28 polyclonal antibody (Santa Cruz Biotechnology) or mouse monoclonal GAPDH antibody (Thermo Fisher Scientific) were used as loading controls. For detection of phosphorylated ␣Syn, mouse Ser-129 phospho-specific antibody (Wako Chemicals USA, Inc., Richmond, VA) was used.
Quantifications of Western Blots-Pixel density values for Western quantification were obtained from TIFF files generated from digitized x-ray films (Kodak) and analyzed with the ImageJ software (44). Before comparison, sample density values were normalized to the corresponding loading control. The adjusted density values were standardized to the control lane to get fold increase. The significance of differences was calculated using Student's t test or one-way ANOVA with Bonferroni's multiple comparison test. p value Ͻ 0.05 was considered to indicate a significant difference.
Ni 2ϩ -NTA Affinity Chromatography-Ulp1 ts mutant cells carrying GAL1-SNCA integrations and His 6 -tagged Smt3 (His 6 -smt3) were pre-grown in 200 ml of SC medium containing 2% raffinose at 30°C overnight. Total cells harvested by centrifugation were transferred to 2 liters of YEPD liquid medium containing 2% galactose for 12 h of induction. Cells were collected and lysed by 25 ml of 1.85 M NaOH containing 7.5% ␤-mercaptoethanol for 10 min on ice. Protein was precipitated in 25 ml of 50% trichloroacetic acid (TCA) and washed with 100% cold acetone. Proteins were suspended in 25 ml of buffer A (6 M guanidine HCl, 100 mM sodium phosphate, 10 mM Tris/HCl, pH 8.0) and rotated for 1 h at 25°C. The supernatant was cleared by centrifugation; the pH was adjusted to 7.0 by 1 M Tris base and supplemented with imidazole to a final concentration of 20 mM. After equilibration of the His GraviTrap column (GE Healthcare) with 5 ml of buffer A containing 20 mM imidazole, proteins were applied to the column, and the flowthrough fraction was collected for analysis. The column was washed with buffer A supplemented with 20 mM imidazole and then with buffer B (8 M urea, 100 mM sodium phosphate, 10 mM Tris, pH 6.3). Then the column was washed with buffer C (50 mM Tris, pH 8.0, 300 mM NaCl, 20 mM imidazole). Finally, the proteins were eluted four times with 1 ml of 200 mM imidazole resolved in buffer C. Protein concentration in the eluted fractions was determined with Bradford assay.
Promoter Shutoff Assays and Chemical Treatments-Yeast cells carrying ␣Syn were pre-grown in SC selective medium containing 2% raffinose overnight and then shifted to 2% galactose SC selective medium to induce the ␣Syn expression for 5 h. Then cells were shifted to SC medium supplemented with 2% glucose to shut off the promoter. At several time points after promoter shutoff, cells were visualized by fluorescence microscopy. For experiments with temperature-sensitive yeast strain smt3 ts , preincubation was performed at 25°C. Induction of ␣Syn expression and the promoter shutoff assay were performed at 25 and 30°C. The reduction of the number of cells displaying ␣Syn inclusions was recorded and plotted on a graph. Drugs used in this study were carbobenzoxyl-leucinylleucinyl-leucinal (MG132) dissolved in dimethyl sulfoxide (DMSO) and phenylmethanesulfonyl fluoride (PMSF) in ethanol (EtOH). Drug treatment was conducted concomitantly with the shift to glucose-supplemented medium in promoter shutoff assays. PMSF was added to the cell suspension to a final concentration of 1 mM. An equal volume of ethanol was added to the cells as a control (45). MG132 treatment was performed as described previously (46). MG132 was applied to the cell suspension in a final concentration of 75 M, and in parallel, an equal volume of DMSO was added to the cells as a control.
Immunoprecipitation-100 g of protein purified by Ni 2ϩ -NTA was incubated with primary antibody (ubiquitin mouse monoclonal antibody, Millipore) at 4°C for 2 h in Immunoprecipitation (IP) buffer (50 mM Tris/HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA) with freshly added 6 mM protease inhibitor mixture (Roche Applied Science), 2 mM DTT, 0.1% phosphatase inhibitor (Roche Applied Science) and then incubated with prewashed protein A-Sepharose beads in IP buffer overnight at 4°C. Beads were washed three times with ice-cold IP buffer; the immunoprecipitated protein was dissolved from the beads by heating in 1ϫ sample loading buffer at 95°C for 10 min, and the samples were subjected to Western blot analyses using rabbit ␣Syn polyclonal antibody (Santa Cruz Biotechnology).
Southern Hybridization and Copy Number Determination-Several transformants were analyzed by Southern hybridization for verification of the integration of ␣Syn-GFP construct into the mutated genomic ura3-1 locus. Isolation of genomic DNA from S. cerevisiae was performed according to standard procedures. 10 g of genomic DNA were subjected to restriction digestion with HindIII. The restriction fragments were resolved on a 1% agarose gel, transferred to a nitrocellulose membrane, cross-linked by UV irradiation for 5 min, and hybridized to a URA3 gene fragment probe. Copy numbers of the integrated vector were estimated according to the profile of the restriction fragments. One copy corresponded to 2.7 ϩ 4.7 kb and two copies to 2.7 ϩ 4.7 ϩ 6.2 kb. For integration of ␣Syn-GFP into the mutated genomic trp1-1 locus, 10 g of genomic DNA were subjected to restriction digestion with EcoRI. One copy of the integrated vector corresponded to 1.9 ϩ 4.2-kb restriction digestion fragments and two copies to 1.9 ϩ 4.2 ϩ 4.6 kb.

RESULTS
␣Syn Is Sumoylated in Yeast-First, we analyzed whether ␣Syn expressed in yeast cells is sumoylated. A temperaturesensitive strain of S. cerevisiae was used with a conditional defect in a gene for an isopeptidase for SUMO deconjugation in a temperature-sensitive manner (ulp1 ts ) (47). Genes for wildtype (WT) ␣Syn or the familial mutant A30P and the His 6tagged yeast SUMO protein Smt3 were integrated into the genome and co-expressed. Down-regulation of the gene for the ULP1 protease activity at the nonpermissive 30°C resulted in an enrichment of SUMO-conjugated proteins (Fig. 1A). SUMO targets were isolated by Ni 2ϩ affinity chromatography under denaturing conditions. The SUMO-modified protein with a molecular mass of ϳ35 kDa can be separated from unmodified 17-kDa ␣Syn. Immunoblotting analysis with a monoclonal antibody against ␣Syn revealed significant sumoylation of both ␣Syn variants (Fig. 1B).
Next, we examined the effect of sumoylation on ␣Syn yeast cells defective in the SUMO-encoding gene (smt3 ts ) (48). The smt3-331 allele expresses a temperature-sensitive Smt3 mutant protein. The mutant Smt3 is dysfunctional at the restrictive temperature (30°C), rendering the protein misfolded (49). The level of SUMO conjugates in the smt3 ts strain was not changed at the nonpermissive temperature, presumably due to the accumulation of misfolded SUMO (Fig. 1C). This is supported by the earlier finding that the phenotype of smt3-331 can be suppressed by WSS1, which had been originally identified as a high copy number suppressor of the temperature-sensitive smt3-331 allele (49). The WSS1 protein acts as SUMO-dependent isopeptidase (50) and presumably detaches misfolded SUMO chains that are caused by the smt3-331 mutation. Consistently, misfolded SUMO chains are accumulated from SUMO conjugates at the nonpermissive temperature.
We previously showed that expression of WT ␣Syn from two copies is under the threshold for yeast growth inhibition (38). Thus, yeast smt3 ts strains expressing WT ␣Syn from two genomically integrated gene copies were constructed, and the number of integrated copies was verified by Southern hybridization. Colony growth was compared in spotting assays between yeast cells with the GAL1 promoter-driven ␣Syn expression under inducing (galactose) or noninducing (glucose) conditions. All strains grew equally well at the permissive temperature (25°C) when sumoylation was not impaired. Expression of WT ␣Syn resulted in growth inhibition in comparison with cells expressing GFP as a control, when sumoylation was down-regulated at the restrictive temperature (30°C). Similar results were obtained for A30P, where high copy plasmid expression normally does not impair yeast growth, whereas defects in sumoylation resulted in a drastic growth inhibition ("ϩSUMO" versus "ϪSUMO" cells depleted of functional SUMO-conjugates in Fig. 1D). This suggests a protective role of the SUMO modification for ␣Syn-expressing yeast cells.
We then assessed whether ␣Syn-mediated cytotoxicity was related to the formation of ␣Syn inclusions. Inclusion formation was followed by live-cell imaging using GFP as a reporter. Quantification of the number of cells displaying fluorescent foci revealed significant increases in cells displaying ␣Syn foci in the absence of sumoylation (Fig. 1E). Control experiments with wild-type yeast (W303) excluded that the difference in the number of cells with inclusions is due to the temperature shift (Fig. 1F).
These results illustrate that inhibition of sumoylation in yeast has a strong growth impact on cells expressing ␣Syn. Growth impairment correlates with the increase of intracellular accumulation of WT ␣Syn or A30P ␣Syn fluorescent foci. This supports that sumoylation protects yeast growth by inhibiting inclusion formation of ␣Syn.

Protective Function of SUMO Requires Direct Modification of ␣Syn at Acceptor Sites That Are Conserved in Eukaryotic Cells-
The protective function of SUMO could be due to a direct sumoylation of ␣Syn or due to an indirect effect through another SUMO target protein. Lys-96 and Lys-102 were mapped as major ␣Syn SUMO acceptor sites in higher cells (15,29). Thus, we then analyzed whether the SUMO acceptor sites in ␣Syn were also conserved in yeast. Double lysine substitutions, K96R/K102R, of WT and A30P ␣Syn were generated. Ulp1 ts yeast cells carrying His 6 -tagged Smt3 were transformed with the double lysine mutants, and SUMO conjugates were purified by Ni 2ϩ -NTA pulldown as above. Considerable amounts of probe were loaded on the gel to increase the detection of low signals. Immunodetection of ␣Syn revealed a significant reduction in sumoylation of the K96R/K102R variant and a complete sumoylation abolishment in the A30P variant carrying these substitutions ( Fig. 2A). These results corroborate that Lys-96 and Lys-102 are conserved as major SUMO acceptor sites of ␣Syn in yeast.
The major sumoylation sites of ␣Syn were replaced in cis to examine whether modification of ␣Syn as direct target of SUMO protects against growth inhibition of yeast. Strains expressing either WT ␣Syn or the K96R/K102R variant from two genomically integrated copies were constructed as above. Expression of the K96R/K102R variant resulted in growth inhibition in contrast to wild-type ␣Syn (Fig. 2B). Fluorescence microscopy studies revealed an increase in the percentage of cells with ␣Syn inclusions for the K96R/K102R mutant (Fig.  2C). These data support that direct sumoylation of ␣Syn at the conserved modification sites (Lys-96 and Lys-102) protects against cytotoxicity and reduces the formation of inclusions.
Phosphorylation of ␣Syn by Human Kinases GRK5 or PLK2 Varies When the Cellular SUMO Pool Is Lowered-␣Syn is phosphorylated predominantly at the residue Ser-129 (18,19).
The effects of phosphorylation on ␣Syn-induced toxicity are complex with reports supporting negative as well as positive impacts on cells (23,24,(51)(52)(53)(54). Therefore, we next investigated the interplay between ␣Syn sumoylation and phosphorylation by examining how changes in sumoylation affect ␣Syn phosphorylation and whether this impacted ␣Syn toxicity. . ␣Syn is sumoylated in S. cerevisiae and impairment of sumoylation increases ␣Syn toxicity and foci formation. A, total protein extract of ulp1 temperature-sensitive yeast cells, defective in SUMO deconjugation, co-expressing integrated ␣Syn (driven by GAL1 promoter at TRP1 locus) and the His 6tagged yeast SUMO protein Smt3 (driven by ADH1 promoter at HIS3 locus). Enriched sumoylated proteins in the ulp1 ts strain in comparison with the control W303 were detected by Western blot with Smt3 antibody (␣-Smt3). EV, yeast cells, transformed with empty vector. B, nickel pulldown of His 6 -tagged yeast SUMO protein Smt3 (His 6 -smt3) in ulp1 ts cells co-expressing ␣Syn. Sumoylated ␣Syn (␣Syn and A30PSyn) was detected in the pulldown fractions with ␣Syn antibody (upper panel). Unmodified ␣Syn was detected in the flow-through. Yeast cells transformed with empty vector were used as a control. Western hybridization of the same blot with Smt3 antibody (lower panel) verified the Ni 2ϩ pulldown (lower panel). C, Western hybridization with Smt3 antibody of total protein extract of smt3 temperature-sensitive yeast cells, co-expressing ␣Syn or empty vector (EV) at permissive (25°C) or restrictive temperature (30°C). D, spotting assay of conditional smt3 ts mutant strain expressing ␣Syn-GFP or A30P-GFP at permissive (25°C; ϩSUMO; Smt3 functional) or restrictive temperature (30°C; ϪSUMO; Smt3 dysfunctional). GAL1-driven ␣Syn-GFP is expressed from two genomically integrated copies. GAL1-driven A30P-GFP is expressed from a 2-m plasmid. GFP, expressed from the same promoter, is used as a control. Yeast cells were spotted in 10-fold dilutions on selection plates containing glucose (␣Syn 'OFF') or galactose (␣Syn 'ON'). E, fluorescence microscopy of smt3 ts cells expressing ␣Syn-GFP or A30P-GFP at permissive (25°C; ϩSUMO) or restrictive temperature (30°C; ϪSUMO). Scale bar, 1 m. F, quantification of the percentage of cells displaying ␣Syn inclusions. W303 cells expressing ␣Syn-GFP or A30P-GFP at 25 or 30°C were used in comparison with smt3 ts cells expressing ␣Syn-GFP or A30P-GFP at permissive (25°C; ϩSUMO) or restrictive temperature (30°C; ϪSUMO). At least 300 cells were counted per strain and experiment. Significance of differences was calculated with t test (***, p Ͻ 0.001, n ϭ 3).
We assessed the combined effects of sumoylation and ␣Syn Ser-129 phosphorylation (␣Syn Ser(P)-129) by overexpressing the human kinases GRK5 or PLK2 from episomal 2-m plasmids in the smt3 ts cells expressing ␣Syn. Heterologous expression of kinases GRK5 or PLK2 resulted in increased phosphorylation of ␣Syn at Ser-129 in comparison with vector control cells (Fig. 3A). Quantification of ␣Syn Ser-129 phosphorylation levels in the presence or absence of functional SUMO revealed a significant increase of ␣Syn Ser-129 phosphorylation when either of the two kinases is expressed (Fig. 3B), whereas expression of GRK5 in the absence of SUMO had a less pronounced effect. Wild-type W303 yeast cells co-expressing K96R/K102R Syn-GFP together with GRK5 or PLK2 were hybridized with ␣Syn Ser(P)-129-specific antibody to examine the Ser-129 phosphorylation level of the sumoylation-deficient ␣Syn variant (Fig. 3C). Quantification of the Western blots revealed a significant increase of the phosphorylation level on Ser-129 of K96R/K102R Syn where the major sumoylation sites of ␣Syn were blocked. Both kinases, GRK5 or PLK2, increased phosphorylation similarly (Fig. 3D).
Expression of GRK5 Alleviates ␣Syn-induced Cytotoxicity and Inclusion Formation in SUMO-deficient Strain-We investigated whether ␣Syn-mediated toxicity was altered when increased ␣Syn Ser-129 phosphorylation levels are combined with functional or dysfunctional SUMO. The effects of GRK5 or PLK2 were tested by spotting assays of smt3 ts cells expressing ␣Syn (Fig. 4A). We found that increased GRK5 suppressed the growth defect associated with impaired sumoylation. Increased PLK2 levels resulted in a less pronounced improvement of growth in comparison with corresponding cells with GRK5 activity (Fig. 4A). The specificity of phosphorylation of GRK5 or PLK2 on Ser-129 was analyzed in greater detail by integrating two copies of an S129A mutant form of ␣Syn in the genome. S129E/S129D mutants were not included in the analysis because recent reports show that they fail to mimic the effect of ␣Syn phosphorylation (9,60). In the presence of functional SUMO, co-expression of S129A with GRK5 had the same growth phenotype as that observed for cells co-expressing WT ␣Syn with GRK5. A slight growth retardation was observed by co-expression of S129A and PLK2 at the permissive temperature. In the absence of functional SUMO, neither kinase could rescue the growth defect of the mutant ␣Syn where the phosphorylation site was missing (Fig. 4A). These data indicate that the SUMO-dependent effect of GRK5 or PLK2 expression on yeast growth depends on the phosphorylation of ␣Syn on Ser-129. We then performed growth assays of cells expressing WT ␣Syn or the K96R/K102R variant to test whether the growth rescue by expression of GRK5 or PLK2 required direct sumoylation of ␣Syn. Co-expression of GRK5 and K96R/ K102R resulted in a striking recovery of growth (Fig. 4B). This suggests that GRK5 directly suppresses a sumoylation defect of ␣Syn. In contrast, expression of PLK2 did not significantly influence yeast growth. This suggests an indirect effect on ␣Syn toxicity caused by down-regulation of the sumoylation activity, which then allows a partial growth recovery by PLK2 expression.
We investigated whether the growth recoveries of smt3 ts cells expressing ␣Syn in the presence of GRK5 or PLK2 are associated with changes in inclusion formation. Phosphorylation of ␣Syn on Ser-129 by GRK5 or PLK2 in sumoylationdeficient cells correlated with decreased accumulation of ␣Syn foci (Fig. 4C). Quantification of the cells displaying ␣Syn inclu-

Sumoylation and Phosphorylation of ␣-Synuclein
NOVEMBER 7, 2014 • VOLUME 289 • NUMBER 45 sions revealed that both GRK5 and PLK2 promote a significant decrease in the percentage of cells bearing fluorescent foci. This effect was Ser-129-dependent, because co-expression of the S129A mutant with either kinase did not reveal decreased accumulation of ␣Syn foci in the absence of SUMO. These results suggest that increased levels of ␣Syn Ser-129 phosphorylation can suppress the ␣Syn-induced cytotoxicity in the SUMO-deficient mutant. PLK2 does not significantly influence yeast growth, although there seems to be a decrease in aggregate formation.
Phosphorylation Promotes Proteasome and Autophagy Degradation of ␣Syn, whereas Sumoylation Preferentially Stimulates Autophagy-We performed GAL1 promoter shutoff experiments and analyzed the impact of blocking these systems by drug treatments to compare the role of proteasome and autophagy-mediated degradation systems on the clearance of ␣Syn inclusions when sumoylation was inhibited. Expression of ␣Syn was induced for 4 h in galactose-containing medium, and the cells were then shifted to glucose-containing medium to repress the promoter. Cells were imaged 2 h after promoter shutoff, and the percentage of cells with inclusions was determined.
Shutoff studies were performed with the mutants ⌬atg1 and ⌬atg7, which render cells unable to perform autophagy. Atg1 is a serine/threonine kinase that acts in autophagy regulation and is essential for autophagy induction (61). Atg7 is an activator of Atg8 and is required for the formation of autophagic bodies (62). Deletion of ATG1 and ATG7 autophagy genes FIGURE 3. Expression of the human kinases GRK5/PLK2 increases ␣Syn Ser-129 phosphorylation in yeast. A, Smt3 ts mutant cells co-expressing ␣Syn and GRK5 or PLK2 at permissive (25°C; ϩSUMO) or restrictive temperature (30°C; ϪSUMO). The phosphorylation level of ␣Syn on Ser-129 was detected by ␣Syn Ser-129 phosphorylation-specific antibody (␣Syn pSer129) when expressed either alone (␣Syn-GFP ϩ empty vector (EV)) or in the presence of GRK5 or PLK2. Immunoblotting analysis of yeast cells expressing S129A-GFP with ␣Syn Ser(P)129 antibody (right panel) was used as a control for antibody specificity. B, quantification of ␣Syn Ser-129 phosphorylation level in the presence or absence of GRK5 and PLK2, respectively, at permissive (25°C; ϩSUMO) or restrictive temperature (30°C; ϪSUMO). Densitometric analysis of the immunodetection of ␣Syn Ser(P)-129 was normalized to the total amount of ␣Syn and relative to ␣Syn ϩ EV at permissive temperature (25°C; ϩSUMO). Significance of differences was calculated with one-way ANOVA with Bonferroni's multiple comparison test (**, p Ͻ 0.01; ##, p Ͻ 0,05 versus empty vector, n ϭ 4). C, Western hybridization of W303 yeast cells co-expressing K96R/K102R-GFP and GRK5 or PLK2. The phosphorylation level of sumoylation-deficient ␣Syn mutant on Ser-129 was visualized with ␣Syn Ser(P)-129. D, quantification of ␣Syn Ser-129 phosphorylation levels of sumoylation-deficient ␣Syn mutant in the presence or absence of GRK5 and PLK2. Densitometric analysis of the immunodetection of ␣Syn Ser(P)-129 was normalized to the total amount of ␣Syn. Significance of differences was calculated with one-way ANOVA (**, p Ͻ 0.01, n ϭ 4).
significantly reduced ␣Syn aggregate clearance 2 h after shutoff. In contrast, cells expressing K96R/K102R ␣Syn cleared inclusions in a similar manner to the isogenic wildtype strain (Fig. 5A).
PMSF as an inhibitor of the autophagy/vacuolar pathway (38) was used in a second approach to study the contribution of autophagy/vacuole for aggregate clearance. PMSF inhibits the activity of numerous vacuolar serine proteases (63) without . Increased ␣Syn Ser-129 phosphorylation level by GRK5/PLK2 expression alleviates the toxicity and reduces foci formation associated with impaired sumoylation in smt3 ts cells. A, spotting assay of smt3 ts cells co-expressing ␣Syn-GFP or S129A-GFP with GRK5 or PLK2 either at permissive (25°C; ϩSUMO) or restrictive temperature (30°C; ϪSUMO). Yeast cells were spotted in 10-fold dilutions on selection plates containing glucose (␣Syn 'OFF'; kinases 'OFF') or galactose (␣Syn 'ON'; kinases 'ON'). B, spotting assay of W303 yeast cells, carrying two copies of GAL1-driven ␣Syn-GFP and K96R/K102R-GFP in the presence of GRK5 and PLK2 or empty vector. Yeast cells were spotted in 10-fold dilutions on selection plates containing glucose (␣Syn 'OFF'; kinases 'OFF') or galactose (␣Syn 'ON'; kinases 'ON'). C, fluorescence microscopy of smt3 ts cells expressing ␣Syn in the presence or absence of GRK5 or PLK2 (left panel). Scale bar, 1 m. Quantification of percentage of cells displaying ␣Syn inclusions in the presence or absence of GRK5/PLK2 (upright) and quantification of cells with foci formation expressing S129A-GFP with and without overexpression of GRK5/PLK2. Significance of differences was calculated with t test with respect to ␣Syn-GFP at the same temperature (**, p Ͻ 0.01, n ϭ 3).

Sumoylation and Phosphorylation of ␣-Synuclein
NOVEMBER 7, 2014 • VOLUME 289 • NUMBER 45 affecting proteasome function (64). PMSF affects autophagic body formation (65) and leads to accumulation of autophagosomes in the vacuole due to decreased degradation of the autophagic bodies (66). Inhibition of autophagic proteases with PMSF resulted in similar impairment in the clearance of inclusions as with the mutant ⌬atg1 strain (Fig. 5B). Cells expressing K96R/K102R ␣Syn cleared inclusions in a similar manner as control cells without drug (ethanol). This suggests that sumoylation supports the autophagy-dependent clearance of ␣Syn.
PMSF was also applied to assess the impact of GRK5 and PLK2 on the clearance of ␣Syn inclusions (Fig. 5B). Expression of GRK5 or PLK2 altered the inclusion clearance significantly and resulted in intermediate levels between WT and K96R/ K102R, suggesting that expression of GRK5 or PLK2 can partially rescue the inclusion clearance through autophagy.
The impact of GRK5 or PLK2 expression and sumoylation on ␣Syn inclusion clearance by the proteasome was analyzed by applying the proteasome inhibitor MG132 (dissolved in DMSO) (67). Quantification of the results of promoter shutoff studies revealed equal inclusion clearance of wild-type ␣Syn in MG132treated cells when compared with the control (DMSO) (Fig. 5C). In contrast, cells were unable to clear inclusions when ␣Syn sumoylation (K96R/K102R) and the proteasome (MG132) were blocked simultaneously. This corroborates that sumoylationdeficient ␣Syn is cleared by the proteasome. Expression of GRK5 in the sumoylation-deficient mutant promoted the proteasome-dependent clearing of inclusions significantly and, accordingly, MG132 treatment resulted in an increased percentage of cells with inclusions. Expression of PLK2 in the sumoylation-deficient mutant could only partially promote proteasomal degradation in comparison with wild-type ␣Syn, suggesting a minor impact on inclusion clearance by the proteasome in comparison with GRK5.
These findings indicate that sumoylated ␣Syn is primarily targeted to the autophagy pathway and nonsumoylated ␣Syn primarily to the proteasome. Inhibition of sumoylation results in inefficient autophagy-mediated aggregate clearance and directs the protein to the proteasome. Expression of the human kinase GRK5 promotes clearance of nonsumoylated ␣Syn to the autophagosome and the proteasome. PLK2 can efficiently phosphorylate nonsumoylated ␣Syn but shows only partial effects on aggregate clearance. This might be due to additional effects on other yet unidentified targets in yeast cells.
Ulp1 SUMO Isopeptidase Activity Increases ␣Syn Inclusion Formation-Decreased sumoylation of ␣Syn impairs inclusion clearance. Therefore, we analyzed whether increased sumoylation of ␣Syn affected the process of inclusion formation and clearance. Expression of ␣Syn-GFP or K96R/K102R ␣Syn in ulp1 ts strain, deficient for SUMO de-conjugation, revealed a general decrease in inclusion formation in comparison with W303, suggesting that the loss of a general SUMO isopeptidase might have multiple effects on the cell (Figs. 1F and 6A). The growth of ␣Syn-GFPand K96R/K102R ␣Syn-expressing cells FIGURE 5. ␣Syn aggregate clearance upon promoter shutoff. A, inhibition of autophagy by deletion of ATG1 and ATG7. Expression of ␣Syn-GFP and K96R/K102R-GFP was induced for 4 h in galactose medium and then the cells were shifted to glucose medium. Quantification of the reduction of inclusions was done 2 h after the promoter shutoff and was presented as the ratio of aggregate clearance in the deletion strain to aggregate clearance in the isogenic wild-type strain. Significance of differences was calculated with t test (* *, p Ͻ 0.01, n ϭ 3). B, inhibition of the vacuolar degradation pathway by PMSF. Quantification was of cells expressing ␣Syn-GFP, K96R/K102R-GFP, and K96R/K102R-GFP and co-expressing GRK5 or PLK2, respectively. ␣Syn-GFP and K96R/K102R-GFP were expressed from two genomically integrated copies. After 4 h of induction of the protein expression in galactose medium, cells were shifted to glucose medium supplemented with 1 mM PMSF dissolved in ethanol (EtOH) or only EtOH as a control. Quantification of the reduction of inclusions was done 2 h after the promoter shutoff. Cells with inclusions were counted and presented as a ratio to the control (EtOH). C, inhibition of the proteasome with MG132. The protein expression was induced as above, and the cells were shifted to glucose medium supplemented with 75 M MG132 and dissolved in DMSO or only DMSO as a control. Quantification of the reduction of inclusions was done 2 h after the promoter shutoff. Cells with inclusions were counted and presented as a ratio to the control (DMSO). Sig-nificance of differences was calculated with one-way ANOVA with Bonferroni's multiple comparison test (**, p Ͻ 0.01; ***, p Ͻ 0.001; n ϭ 3; #, p Ͻ 0.05 versus K96R/K102R-GFP) (Bonferroni's multiple comparison test).
in ulp1 ts was inhibited similarly (Fig. 6B). The cells showed partial cytoplasmic GFP staining additionally to fluorescent foci (Fig. 6C), whereas expression of ␣Syn-GFP and K96R/ K102R ␣Syn in the W303 background did not reveal any cytoplasmic GFP staining. Promoter shutoff experiments revealed a slower rate of inclusion clearance of the sumoylation-deficient K96R/K102R mutant in comparison with ␣Syn (Fig. 6D). The results suggest that an increase of the pool of sumoylated proteins by inhibition of SUMO de-conjugation can change the inclusion formation and localization of ␣Syn without changing its toxicity. One possible explanation might be that high pools of free SUMO necessary for sumoylation of ␣Syn are required to decrease its toxicity and impact on cell growth.
Phosphorylation Promotes Ubiquitination and Degradation of ␣Syn-Ubiquitination is the common post-translational modification for proteasome-dependent protein degradation and is usually primed by a kinase reaction. We showed that increased levels of ␣Syn phosphorylation on Ser-129 affect the clearance of inclusions by the proteasome. These results prompted us to analyze how sumoylation or phosphorylation influences ubiquitination of ␣Syn. Wild-type cells expressing His 6 -tagged ␣Syn were grown, and the protein was purified from cell extracts using Ni 2ϩ affinity chromatography. Immunoprecipitation was performed with ubiquitin antibodies and revealed that ␣Syn monomers are mono-ubiquitinated in yeast (Fig. 7A), in agreement with findings in higher organisms (68 -70). Phosphoantibodies showed that mono-ubiquitinated ␣Syn was simultaneously phosphorylated on Ser-129 (Fig. 7A).
We next assessed whether ubiquitination of ␣Syn was affected by sumoylation and whether phosphorylation altered ␣Syn ubiquitination. For this, His 6 -tagged ␣Syn was expressed from a 2-m plasmid in smt3 ts cells. The effect of GRK5 and PLK2 was investigated by co-expression of each kinase in the presence or absence of SUMO (Fig. 7B). Ni 2ϩ affinity chromatography of His 6 -tagged ␣Syn was performed, followed by immunoprecipitation of the protein with an antibody against ubiquitin. Immunoblotting analysis revealed different patterns of ubiquitinated ␣Syn species ranging from 22 to 36 kDa (Fig.  7B). In the absence of GRK5 or PLK2, only a single molecular band at around 29 kDa was precipitated. The presence of either kinase resulted in multiple distinct bands, including a major band of 22 kDa. The additional smear pattern of modified ␣Syn to higher molecular weights was especially pronounced when sumoylation was down-regulated. This might be due to monoubiquitination (22 kDa), di-ubiquitination (29 kDa), or tri-ubiquitination, as described earlier (11). The expression of GRK5 resulted in larger effects on the ubiquitination of ␣Syn than those observed with PLK2, especially in the absence of SUMO. This was consistent with the stronger suppression of ␣Syn toxicity by GRK5 when SUMO was down-regulated, in comparison with PLK2.
GAL1 promoter shutoff assays were performed to determine the effect of sumoylation and increased Ser-129 phosphorylation by GRK5 and PLK2 on ␣Syn stability. As described above, the promoter was shut off after 4 h, and cells were collected at various time points. Immunoblotting analysis revealed a reduction in the level of ␣Syn with time. Phosphorylation of ␣Syn by GRK5 or PLK2 resulted in a slight decrease of the protein levels after 18 h in comparison with the control (Fig. 8A). We assessed the role of the proteasome and autophagy-mediated degradation systems on ␣Syn stability and analyzed the impact of blocking these systems by drug treatments. Inhibition of the proteasome with MG132 had slight impact on the stability of ␣Syn after 18 h, whereas inhibition of the vacuolar/autophagy path- way with PMSF resulted in increased protein stability (Fig. 8B). These results corroborate our previous findings (38). Western blot hybridization of MG132-treated cells with ubiquitin antibody confirmed the effectiveness of MG132 as a proteasome inhibitor (Fig. 9).
We analyzed whether direct inhibition of ␣Syn sumoylation, by blocking the major sumoylation sites (K96R/K102R), affected the steady state of protein stability. GAL1 promoter shutoff experiments revealed that the SUMO-deficient K96R/ K102R ␣Syn variant is a highly stable protein (Fig. 8C). We next tested whether increased ␣Syn phosphorylation by GRK5 and PLK2 could alter the protein stability of the SUMO-deficient K96R/K102R variant. Immunoblotting analysis after promoter shutoff revealed a significant reduction in the levels of ␣Syn. Phosphorylation of ␣Syn by GRK5 or PLK2 therefore destabilizes ␣Syn significantly when sumoylation is impaired (Fig. 8, D  and E). Inhibition of the proteasome with MG132 and vacuolar/ autophagy pathway with PMSF resulted in a significant increase in protein stability. The results support the data from aggregate clearance assays and suggest that GRK5 and PLK2 affect the protein stability of the SUMO-deficient mutant by directing the protein to the vacuole and proteasome for degradation.
We examined whether the effect depends directly on phosphorylation of Ser-129. GAL1 promoter shutoff assays were performed with S129A ␣Syn mutant in smt3 ts strain at permissive (ϩSUMO) or restrictive temperature (ϪSUMO). S129A-␣Syn mutant revealed a decrease in protein level after 18 h of promoter shutoff (Fig. 10). However, the mutant protein was much more stable than wild-type ␣Syn (Fig. 8A). Down-regulation of sumoylation at a restrictive temperature resulted in no significant decrease of the protein stability. The data suggest that phosphorylation of ␣Syn at Ser-129 decreases the protein stability, which is further affected by sumoylation. These data corroborate the results with the sumoylation-deficient K96R/ K102R mutant. Expression of GRK5 or PLK2 did not affect the stability of S129A-␣Syn. The results indicate that the effect of GRK5 and PLK2 expression on ␣Syn protein stability depends directly on Ser-129. FIGURE 7. ␣Syn is ubiquitinated in yeast cells. A, ␣Syn-His 6 protein was purified by Ni 2ϩ pulldown and subjected to immunoprecipitation with ubiquitin antibody. The ubiquitinated and phosphorylated ␣Syn was detected by ␣Syn and ␣Syn Ser(P)-129-specific antibody, respectively. Empty vector (EV) was used as a control. B, Smt3 ts cells expressing ␣Syn-His 6 co-transformed with GRK5 or PLK2 and empty vector of the kinases (EV) as a control at permissive (25°C; ϩSUMO) or restrictive temperature (30°C; ϪSUMO). The purified ␣Syn protein from Ni 2ϩ pulldown was subjected to ubiquitin immunoprecipitation (IP Ubi). As a control, the same experiments were performed without addition of ubiquitin antibody. The ubiquitinated ␣Syn was analyzed by Western hybridization with an antibody against ␣Syn. Western hybridization of the same blots after stripping with ubiquitin antibody (lower panels). A representative result is shown from three independent experiments. FIGURE 8. Effect of sumoylation and increased ␣Syn Ser-129 phosphorylation by GRK5/PLK2 on ␣Syn protein stability. GAL1 promoter shutoff studies and drug treatments. A, Smt3 ts yeast cells expressing ␣Syn with or without GRK5 or PLK2 at permissive temperature (25°C) were induced for 4 h in galactose (␣Syn on) and then transferred to glucose containing medium (␣Syn off). Immunoblotting analysis was performed at the indicated time points after promoter shutoff with ␣Syn antibody and GAPDH antibody as loading control. B, W303 cells expressing ␣Syn, K96R/K102R (C), K96R/K102R ϩ GRK5 (D), or K96R/ K102R ϩ PLK2 (E) were induced for 4 h in galactose (␣Syn on) and then transferred to glucose containing medium (␣Syn off). The glucose medium was supplemented with 75 M MG132 or 1 mM PMSF. Immunoblotting analysis was performed at the indicated time points after promoter shutoff with ␣Syn antibody and GAPDH antibody as loading control. A representative result is shown from three independent experiments. Right panels, densitometric analysis of the immunodetection of ␣Syn-GFP relative to the GAPDH loading control. Significance of differences was calculated with one-way ANOVA with Bonferroni's multiple comparison test (***, p Ͻ 0.001; **, p Ͻ 0.01; *, p Ͻ 0.05; ##, p Ͻ 0.05 versus 0 h (Bonferroni's multiple comparison test)).
These results demonstrate that phosphorylation at Ser-129 promotes ␣Syn ubiquitination and decreases its stability. The data support a complex cross-talk between sumoylation-and phosphorylation-mediated ubiquitination of ␣Syn.

DISCUSSION
Here, we used S. cerevisiae as a model to investigate the molecular interplay between sumoylation and phosphorylation in the clearance of ␣Syn (summarized in Fig. 11). We uncovered a complex cross-talk between these post-translational modifications that impact ubiquitination and thereby influence the degradation of ␣Syn by autophagy or the 26 S proteasome. Ultimately, the differential processing of ␣Syn by these systems interferes with inclusion formation and cytotoxicity.
␣Syn undergoes numerous post-translational modifications such as phosphorylation, ubiquitination, nitration, acetylation, O-glycosylation, and sumoylation. ␣Syn was found to be a SUMO target in cultured cells and in a rat animal model of PD (15,29), but the number of sumoylation studies of ␣Syn is very limited in comparison with those on other post-translational modification publications, limiting our understanding of the implications of sumoylation on ␣Syn biology. Here, we showed that both WT ␣Syn as well as the A30P mutant are sumoylated in vivo in yeast at Lys-96 and Lys-102, two sumoylation sites that are conserved in eukaryotes (15,29). By decreasing the cellular SUMO pool, or by mutating the codons for the major SUMO sites, we determined that sumoylation protects yeast cells against ␣Syn-mediated cytotoxicity and inclusion formation. Previously, sumoylation was suggested to keep ␣Syn in   10. Effect of sumoylation and GRK5/PLK2 expression on S129A-GFP protein stability. GAL1 promoter shutoff studies. A, Smt3 ts yeast cells expressing S129A-GFP with or without GRK5 or PLK2 at permissive temperature (25°C, ϩSUMO) and restrictive temperature (30°C, ϪSUMO) were induced for 4 h in galactose (␣Syn on) and then transferred to glucose-containing medium (␣Syn off). Immunoblotting analysis was performed at the indicated time points after promoter shutoff with ␣Syn antibody and GAPDH antibody as loading control. A representative result is shown from three independent experiments. B, densitometric analysis of the immunodetection of S129A-GFP relative to the GAPDH loading control. Significance of differences was calculated with one-way ANOVA with Bonferroni's multiple comparison test (**, p Ͻ 0.01; ##, p Ͻ 0.05 versus 0 h (Bonferroni's multiple comparison test).
solution and, therefore, decrease ␣Syn aggregation (15). Similarly, sumoylation was found to modulate the solubility of mutant huntingtin, Ataxin 7, androgen receptor, and STAT1, also reducing the toxicity of these proteins in other degenerative diseases (71)(72)(73). Consistently, impairment of sumoylation in yeast resulted in a significant increase in the number of cells displaying ␣Syn inclusions. This further supports the beneficial regulatory role of sumoylation in inhibiting ␣Syn inclusion formation in vivo.
Sumoylation and phosphorylation are both reversible dynamic processes, which can actively interfere with each other and modulate the molecular features of their substrates. The major ␣Syn phosphorylation site at Ser-129 is also used in yeast when human kinases are expressed (24). Here, we expressed and analyzed the effects of two of the most efficient human kinases in yeast, i.e. the G protein-coupled kinase GRK5 (21,25) and the Polo-like kinase PLK2 (26,28). Our study revealed a significant increase of ␣Syn phosphorylation at Ser-129 in yeast in the presence of human PLK2 or GRK5 kinases. PLK2 phosphorylated ␣Syn with similar efficiency in the presence or absence of functional SUMO. GRK5 phosphorylated preferentially ␣Syn in cells with an intact sumoylation machinery. The difference in the substrate specificity suggests that other mechanisms or phosphorylation of residues other than Ser-129 could facilitate ␣Syn clearance by overexpression of GRK5.
Several examples have been reported in the literature where phosphorylation depends on the sumoylation profile of target proteins (74,75). Sumoylation can modulate the specific interaction with kinases or phosphatases by changing substrate surfaces and activity. In particular, sumoylation of protein-tyro-sine phosphatase 1B has been shown to reduce catalytic activity and therefore change the phosphorylation status of substrates (76).
Accumulating evidence suggests that ␣Syn post-translational modifications modulate ␣Syn-mediated toxicity and aggregate formation (9,15,20,23,29,68,69,77). However, there is still no consensus of the effects of different modifications on ␣Syn aggregation and toxicity (15,23,24,54,78). Although earlier studies did not observe the effects of ␣Syn phosphorylation at Ser-129 on ␣Syn-mediated toxicity and aggregation (54,79), protective roles of ␣Syn Ser-129 phosphorylation were described in a strain-specific manner in yeast. Therefore, the specific genetic context was proposed to determine the sensitivity to changes in ␣Syn phosphorylation (80). This suggests a complex and subtle cross-talk between different modifications that can change features of the target protein, including inclusion formation, stability, and the affinity to the autophagic or the proteasome degradation pathways. Here, we focused on the interplay between ␣Syn sumoylation and Ser-129 phosphorylation. Increased ␣Syn Ser-129 phosphorylation induced by GRK5 can rescue yeast cells from ␣Syn-mediated cytotoxicity associated with sumoylation impairment. Alleviation of ␣Syn-mediated cytotoxicity in SUMO-deficient cells correlates with a decreased number of cells presenting ␣Syn intracellular inclusions. Expression of GRK5 induced a strong improvement on yeast growth when the sumoylation was impaired both in trans and in cis. We found that PLK2 might cause additional effects in yeast, in agreement with a recent study where we reported a specific role of PLK2 on ␣Syn inclusion formation and toxicity in yeast independent of the level of ␣Syn phosphorylation on Ser-129 (24).
The dynamic process of the ␣Syn aggregate formation depends on the equilibrium between synthesis and degradation, which determines the protein levels of ␣Syn. An important question is how ␣Syn degradation is distributed between the ubiquitin-proteasome system and the autophagy-lysosome/vacuole pathway (69). At low levels, ␣Syn seems to be preferentially degraded by the ubiquitin-proteasome system, whereas increased ␣Syn expression stimulates autophagy as the main degradation pathway (34). We previously found that autophagy represents the major pathway for aggregate clearance in yeast after the shutoff of further protein biosynthesis, allowing cells to recover from ␣Syn toxicity (38).
One of the major findings of this study is that sumoylation of ␣Syn promotes aggregate clearance by autophagy. ␣Syn clearance is impaired when sumoylation is inhibited either by reducing the cellular SUMO pool or by amino acid substitutions of the SUMO target sites of ␣Syn. Another major finding is that phosphorylation of ␣Syn by GRK5 can compensate for this effect. The protective role of PLK2, which can form a complex with ␣Syn and can also induce the autophagy pathway (23), seems to be more complicated and might include additional phosphorylation target proteins. The discrepancy between a clear PLK2 effect on inclusion formation and only a mild protective effect on yeast growth suggests that cellular survival does not only depend on inclusion clearance but requires additional protection pathways. FIGURE 11. ␣Syn clearance and degradation in yeast. Proteasome and autophagy/vacuole as major degradation pathways are depicted. When synthesis of ␣Syn is switched off, wild-type yeast cells clear ␣Syn aggregates within hours and regain normal growth rates (38). In the presence of functional SUMO, the aggregates are primarily cleared by the autophagy/vacuolar pathway. When sumoylation is impaired, the aggregate clearance through autophagy/vacuolar pathway is prevented, and the proteasomal degradation is promoted. Increase of Ser-129 phosphorylation level by GRK5 or PLK2 rescues the autophagic aggregate clearance and additionally promotes the proteasomal degradation. For monomers, degradation of soluble ␣Syn monomers occurs through both pathways. Inhibition of ␣Syn sumoylation has a strong effect on monomer protein stability, significantly increasing the half-life of the protein and inhibiting the degradation through both pathways. Phosphorylation at Ser-129 by GRK5 or PLK2 decreases the protein stability and promotes degradation of soluble ␣Syn through proteasome and autophagy pathways.
Sumoylation and phosphorylation are two post-translational modifications of ␣Syn, which protect against ␣Syn-induced toxicity. However, they represent distinct signals for the processing of ␣Syn by different degradation pathways. Whereas sumoylation primarily targets ␣Syn for autophagy, phosphorylation by kinases such as GRK5 has a dual effect because it partially rescues the autophagy pathway but also promotes increased ubiquitination and a reduced half-life of the protein.
Phosphorylation is a well known priming reaction for ubiquitination (11,81), and our data suggest that increased phosphorylation of ␣Syn presumably results in increased ubiquitination and proteasome-mediated degradation. Proteasome inhibitor studies further support that the phosphorylation-dependent degradation of ␣Syn is promoted by the proteasome. A dual modification that is interdependent allows a subtle fine-tuning as a molecular mechanism to selectively control ␣Syn turnover in response to sumoylation or phosphorylation input signals. Sumoylation might induce structural and conformational changes in ␣Syn and thus modulate the interaction with different kinases, which have various effects in the channeling to distinct degradation pathways.
Our study provides evidence, for the first time, that the degree of switching between autophagic and proteasomal degradation of ␣Syn is linked to a molecular cross-talk between sumoylation and phosphorylation. Sumoylation preferentially directs ␣Syn aggregates toward autophagy, and phosphorylation can shift the fate of ␣Syn to increased ubiquitination and proteasome degradation. Ultimately, a deeper understanding of this cross-talk will enable the design of effective strategies for directing ␣Syn for processing by the desired degradation machinery and may therefore constitute the basis for novel therapeutic strategies in PD and other synucleinopathies.