Alpha-synuclein functions in the nucleus to protect against hydroxyurea-induced replication stress in yeast

Abstract

Hydroxyurea (HU) inhibits ribonucleotide reductase (RNR), which catalyzes the rate-limiting synthesis of deoxyribonucleotides for DNA replication. HU is used to treat HIV, sickle-cell anemia and some cancers. We found that, compared with vector control cells, low levels of alpha-synuclein (α-syn) protect S. cerevisiae cells from the growth inhibition and reactive oxygen species (ROS) accumulation induced by HU. Analysis of this effect using different α-syn mutants revealed that the α-syn protein functions in the nucleus and not the cytoplasm to modulate S-phase checkpoint responses: α-syn up-regulates histone acetylation and RNR levels, maintains helicase minichromosome maintenance protein complexes (Mcm2–7) on chromatin and inhibits HU-induced ROS accumulation. Strikingly, when residues 2–10 or 96–140 are deleted, this protective function of α-syn in the nucleus is abolished. Understanding the mechanism by which α-syn protects against HU could expand our knowledge of the normal function of this neuronal protein.

INTRODUCTION

Progressive degeneration of dopaminergic neurons in a region of the midbrain called the substantia nigra pars compacta causes slowness of movement, resting tremor, rigidity and postural instability—the symptoms of Parkinson's disease (1–3). The pathological hallmark of this disease is the accumulation of cytoplasmic inclusions in affected neurons, and the principal component of these inclusions is the protein α-synuclein (α-syn) (4). Human molecular genetic studies subsequently uncovered that missense mutations (5–7) and duplications (8) in the α-syn locus cause early-onset PD. Although several other loci have been linked to PD, α-syn is thought to be the most important one regarding sporadic PD.

α-Syn is an abundant neuronal protein, of uncertain function, that contains 140 amino acids and is composed of three regions. The N-terminal residues 1–60 can adopt an α-helical conformation which avidly binds to membrane phospholipids (9–11). The NAC domain, residues 61–95, is a hydrophobic segment that promotes aggregation and fiber formation (10,12). The acidic C-terminal residues 96–140 control nuclear localization, exhibit chaperone-like functions (13) and are subject to phosphorylation at serine and tyrosine residues (14–16). Depending on the cellular milieu, α-syn can be unfolded, α-helical or β-sheet (10,17,18). Normal levels of α-syn are thought to control synaptic vesicle recycling and SNARE complex assembly (13,19), whereas elevated levels cause vesicle trafficking defects (19,20), proteasome dysfunction (21,22), reactive oxygen species (ROS) production (23,24) and damage to cellular membranes (25,26).

Protective functions of α-syn have been discovered and new ones are emerging. For example, α-Syn protects cells from the mitochondrial toxins paraquat and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (27–29). siRNA-mediated down-regulated expression of α-syn in the rat brain causes rapid neurodegeneration (30), which implies a protective function. α-Syn exhibits a non-classical chaperone activity that maintains SNARE-complex assembly in presynaptic nerve terminals during aging (13,31). α-Syn may inhibit apoptosis in neurons through covalent hetero-oligomerization of cytochrome c (32), and α-syn exhibits ferrireductase activity (33) which may help cells maintain the Fe(II)/Fe(III) ratio required for synthesis of dopamine by the Fe(II)-dependent dopamine hydroxylase. If α-syn aggregates and fibrilizes, the concentration of the protective monomer could be drastically decreased, resulting in a loss of protective functions. In this study, which uses yeast as a model for PD (34), we found a new protective function for α-syn.

RESULTS

High levels of α-syn inhibit growth and cause ROS to accumulate

Strains and plasmids are given in Tables 1 and 2. Wild-type alpha-synuclein is referred to as α-syn. Cells transformed with the 2 μ pAG426-α-syn (HiTox) plasmid expressed ∼4-fold more α-syn than cells transformed with the 2 μ pESC-α-syn (LoTox) plasmid (Fig. 1A). The HiTox α-syn plasmid (34) caused much more growth inhibition and ROS accumulation than the LoTox α-syn plasmid (Fig. 1B and C). Consequently, to avoid the toxicity due to high expression, the LoTox α-syn plasmid was used in many, but not all, of the experiments described here.

Table 1.

Strains used in this study

Name	Genotype	Source
ɛ1278b	Matα ura3-52 his3	Roberts et al. (35)
ɛ1278b-Pom34-RFP	Matα Pom34-RFP::HIS3 ura3-52 his3	This study
KT3251	Matα leu2 his3 ura3-52 Pom34-mCherry RFP::HIS3	Tatchell et al. (38)
W303	Matα leu2-3,112 ade2-1 can1-100 his3-11,15 ura3-1 trp1-1 RAD5	Zhao et al. (78)
W303-mec1▵sml1▵	Mata mec1▵::TRP1 sml1▵::HIS3 RAD5 in W303	Zhao et al. (78)
W303-rad53▵sml1▵	Mata sml1▵::URA3 rad53▵::HIS3 RAD5 in W303	Zhao et al. (78)
X1215-6B	Mata CFP-Pol30 Fob1-YFP Mcm2-RFP TRP1 lys2▵ in W303	Xiaolan Zhao
X164-14A	Matα mec1-21 in W303	Xiaolan Zhao
S288C	MATa his3▵1 leu2▵0 met15▵0 ura3▵0	This study
S288C-dia2▵	MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 dia2▵::Kanr	This study
S288C-mbp1▵	MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 mbp1▵::Kanr	This study
S288C-not4▵	MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 not4▵::Kanr	This study
S288C-rtt109▵	MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 rtt109▵::Kanr	This study
S288C-vps75▵	MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 vps75▵::Kanr	This study
S288C-RNR1-TAP	MATa his3▵1 leu2▵0 met15▵0 ura3▵0,RNR1-TAP-HIS3	This study
S288C-RNR3-TAP	MATa his3▵1 leu2▵0 met15▵0 ura3▵0,RNR3-TAP-HIS3	This study
S288C-RNR4-TAP	MATa his3▵1 leu2▵0 met15▵0 ura3▵0,RNR4-TAP-HIS3	This study

Table 2.

Plasmids used in this study

Name	Description	Source
LoTox plasmidsa
pESC-URA	pESC-URA-GAL	This study
pESC-α-syn	pESC-URA-GAL-α-syn	This study
pESC-A30P	pESC-URA-GAL-A30P	This study
pESC-A53T	pESC-URA-GAL-A53T	This study
pESC-S129A	pESC-URA-GAL-S129A	This study
pESC-S129D	pESC-URA-GAL-S129D	This study
pESC-α-syn-▵N	pESC-URA-GAL-α-syn(▵2-10)	This study
pESC-A30P-▵N	pESC-URA-GAL-A30P(▵2-10)	This study
pESC-α-syn-▵C	pESC-URA-GAL-α-syn(▵96-140)	This study
pESC-α-syn-NES	pESC-URA-GAL-α-syn-NES	This study
pTF305b	pRS316-GAL-GFP-α-syn	Flower et al. (74)
HiTox plasmidsa
pAG426-α-syn	pAG426-GAL-α-syn	This study
pAG426-A30P	pAG426-GAL-A30P	This study
pAG426-A30P-▵N	pAG426-GAL-A30P(▵2-10)	This study
pAG426-α-syn-▵N-GFP	pAG426-GAL-α-syn(▵2-10)-EGFP	This study
pAG426-GFP-α-syn	pAG426-GAL-EGFP-α-syn	This study
pAG426-GFP-α-syn-▵C	pAG426-GAL-EGFP-α-syn-(▵96-140)	This study
pAG426-GFP-α-syn-NES	pAG426-GAL-EGFP-α-syn-NES	This study

^a2 μ plasmid.

^bCentromeric plasmid.

Figure 1.

α-Syn suppresses HU toxicity. (A) α-Syn expression from LoTox and HiTox plasmids. Cells transformed with the indicated plasmids were induced for 3 h and then the lysates were subjected to SDS–PAGE and western blotting. Strain, ɛ1278b (A–F). Pgk, loading control. (B) Growth properties of cells transformed with LoTox or HiTox plasmids. Serial dilutions (1:10) were spotted onto SD and SG plates, and the plates were incubated at 30°C for 2–3 days. SD, synthetic dextrose non-inducing media; SG, synthetic galactose-inducing media. (C) ROS accumulation. Cells transformed with the indicated plasmids were induced for 3 days and then stained with DCFH-DA for 1 h (see Methods). Images were obtained using differential interference contrast (DIC) and fluorescence microscopy (DCF). The bottom panel gives the percentage of cells (n > 700 per experiment) staining for ROS. Scale bar = 5 μm. (D) α-Syn protects against HU. Cells transformed with LoTox plasmids were serially diluted and spotted onto SD, SG and SG + HU plates, which were incubated at 30°C for 3–5 days. (E) Western blot of α-syn and mutants. Cells were induced for 3 h. Plasmids: pESC-A30P, A53T, S129A and S129D. (F) α-Syn is expressed after long-term HU treatment. Cells collected from 1-week-old spot assay plates (200 mm HU) were pre-grown, transferred into inducing media for 3 h, lysed and subjected to western blotting. Plasmid: pESC-α-syn. (G) α-Syn suppresses HU sensitivity in the indicated strains. The growth assay was conducted as described in (B). Plasmids: pESC-α-syn and empty vector.

Low levels of α-syn protect cells from hydroxyurea

We asked whether low levels of α-syn might modulate the cellular response to genotoxins, and found that α-syn either has no effect or enhances the toxicity of all test compounds (Supplementary Material, Fig. S1), except for one. Hydroxyurea (HU), which blocks replication by inhibiting ribonucleotide reductase (RNR) and is used to treat several diseases, is the exception. Probing this effect in more detail, we discovered that even several LoTox α-syn mutants (A30P, A53T, S129A and S129D), two of which (A30P and A53T) are pathogenic, also protect cells from the growth inhibitory properties of HU (200 mm) (Fig. 1D). Western blotting confirmed that cells expressed the various mutants, and cells expressed α-syn even after 7 days of incubation with HU (Fig. 1E and F). The strongest protection against HU occurred in ɛ1278b (35,36) followed by BY4741 and W303 wild-type strains (Fig. 1G).

Evidence that α-syn functions in the nucleus to protect cells from HU

The following α-syn mutants were used to probe the protection against HU. First, α-syn(Δ2-11), which lacks α-helical structure and fails to bind to membranes, is non-toxic in the yeast model (10). A minor change was made in that α-syn(Δ2-10) was used instead of α-syn(Δ2-11). Second, α-syn(Δ96-140) fails to enter the nucleus in human cells (16), and α-syn-NES fails to enter the nucleus because of the attached nuclear exclusion sequence (37). These two mutants were used to test the importance of nuclear localization. In the following experiments, α-syn(Δ2-10) and A30P(Δ2-10) are referred to as α-syn-ΔN and A30P-ΔN, respectively, and α-syn(Δ96-140) as α-syn-ΔC.

The localization of α-syn-ΔN, α-syn-ΔC and α-syn-NES was assessed using fluorescence microscopy in conjunction with green fluorescence protein (GFP) tags. To visualize the nucleus, we used a strain containing an integrated red fluorescent protein (RFP)-labeled nuclear membrane protein (Pom34-RFP). Nuclear enrichment of α-syn is particularly evident in the cells expressing GFP-α-syn and α-syn-ΔN-GFP (Fig. 2A, top two panels) and less evident in cells expressing α-syn-ΔC-GFP and GFP-α-syn-NES (Fig. 2A, bottom two panels). To quantitatively assess the relative levels of GFP-α-syn in the nucleus and cytoplasm, the nuclear-to-cytoplasmic fluorescence ratio (16,38) was measured for cells expressing each construct. Nuclear enrichment follows the order GFP-α-syn>GFP-α-syn-ΔC>>GFP-α-syn-NES (Fig. 2B). Note that the nuclear enrichment of GFP-α-syn-ΔN is indistinguishable from the full-length protein. Comparing GFP-α-syn to GFP-α-syn-NES, we point out that, at minimum, 40% more GFP-α-syn molecules are in the nucleus than GFP-α-syn-NES molecules.

Figure 2.

α-Syn truncation mutants fail to protect against HU. (A) Localization of GFP tagged α-syn mutants. The Pom34-RFP strain (KT3251) transformed with the indicated GFP-tagged α-syn plasmids (pAG426 plasmids) was induced for 3 h. Pom34-RFP is a marker of the nuclear envelope. GFP and RFP signals were detected by fluorescence microscopy. Arrows indicate nuclei. Scale bar = 5 μm. (B) Plot of the GFP-α-syn nuclear-to-cytoplasmic fluorescence ratio for different α-syn constructs. GFP intensity in the nucleus and cytoplasm was measured by Image J software. The y-axis is the GFP-α-syn nuclear-to-cytoplasmic fluorescence ratio. Data bars are the average ± s.e.m. of three independent experiments (n = 200–250; *P< 0.05; two-tailed Student's t-test). (C and D) α-Syn mutants fail to suppress HU-induced growth defects. Serial dilutions of cells transformed with the indicated plasmids were spotted onto SD, SG and SG + HU plates and incubated at 30°C for 3–4 days. C, plasmids: pESC and pAG426. D, plasmids: pESC. Strain, ɛ1278b (C–E). (E) Western blots of α-syn and mutants. Lysates of cells expressing the various mutants were analyzed by western blotting. Induction time was 5–6 h. E, plasmids: pESC.

Truncation mutants were also evaluated in the growth assay. Figure 2C compares the effects of full-length α-syn and the N-terminal truncations on the growth of ɛ1278b cells treated with HU (200 mm). Cells expressing α-syn or A30P grew much better than vector control cells, whereas cells expressing α-syn-ΔN or A30P-ΔN grew similar to vector control cells. This result shows that α-syn residues 2–10 are required for protection against HU. Figure 2D compares full-length α-syn to the C-terminal truncation mutant. Similarly, cells expressing α-syn grew much better than vector control cells, whereas cells expressing α-syn-ΔN, α-syn-ΔC or α-syn-NES grew similar to vector control cells. Because α-syn truncations and the NES construct were expressed in yeast cells (Fig. 2E), their failure to protect against HU is not due to low expression. Furthermore, because α-syn-NES, which is mostly excluded from the nucleus, fails to protect against HU suggests that full-length α-syn functions in the nucleus to protect against HU.

HU triggers rapid translocation of α-syn into the nucleus

Given that α-syn translocates into the nucleus of mammalian cells in response to oxidative stress (39), we asked whether α-syn does the same in yeast cells in response to HU. The nuclear pore protein Pom34 (tagged with the RFP) was used as a marker of the nuclear membrane. A GFP-α-syn fusion protein was used to monitor α-syn (GFP-α-syn) localization. After a 5 h induction time in inducing media in the absence of HU, α-syn localized primarily to the plasma membrane in early log phase cells (Fig. 3A). In contrast, in identically treated cells with HU (100 mm), α-syn localized to the cytoplasm and nucleus in 15% of early log phase cells (Fig. 3A and C). After a long incubation (24 h; OD₆₀₀ > 3.7), α-syn localized to the cytoplasm and nucleus in about 40% cells (Fig. 3B and C), and no difference existed between HU-treated cells and control cells. The results show that the nuclear localization of α-syn is accelerated by, but does not depend on, HU.

Figure 3.

α-Syn translocates into the nucleus in response to HU. (A and B) Nuclear localization of GFP-α-syn with or without HU treatment. The Pom34-RFP strain (ɛ1278b) transformed with the pTF305 (GFP-α-syn) plasmid was induced for 5 h (A) or 24 h (B) without or with HU (the last 1 h with 100 mm HU). Fluorescence microscopy was used to visualize the RFP- and GFP-tagged proteins. Scale bar = 5 µm. Arrows indicate nuclei. (C) Plot of the percentage of cells with nuclear GFP-α-syn. The percentage of cells with nuclear localized α-syn in (A) and (B) was determined. Data bars are the average ± s.e.m. of 2–3 independent experiments (n = 600–1000; **P< 0.01; two-tailed Student's t-test). (D–F) Immunogold staining of α-syn. Cells (ɛ1278b) transformed with the LoTox α-syn plasmid were induced for 5 h without (D) or with HU (the last 1 h with 100 mm HU) (E). Cells were fixed for transmission electron microscopy (see Materials and Methods). The rectangular area in (E) is enlarged in far right image. N = nucleus; V = vacuole; CW = cell wall. Scale bar = 200 nm. The percentage of gold particles in nucleus with or without HU is shown in (F). Three areas from the membrane, cytoplasm and nucleus were randomly chosen, and gold particles were manually counted using Image J software (see Material and Methods for details). Data bars are the average ± s.e.m. of two independent experiments (n = 30; *, P< 0.05; two-tailed Student's t-test).

To confirm the subcellular localization of α-syn, transmission electron microscopy (TEM) in conjunction with immunogold labeling was used to detect α-syn (Fig. 3D–F). Cells transformed with the LoTox α-syn plasmid were incubated in inducing media for 4 h with or without added HU and then prepared for TEM analysis. In the absence of HU, α-syn-gold particles decorated the perimeter of cells, consistent with membrane localization (Fig. 3D). Less than 4% of the α-syn-gold particles were found in the nucleus (Fig. 3F). In contrast, when the same cells were incubated with HU (last hour with 100 mm) a dramatic change in localization occurred in 10–15% of cells: α-syn-gold particles were distributed throughout the cell instead of being predominantly confined to the plasma membrane (Fig. 3E). Specifically, 40, 38 and 22% of the α-syn-gold particles were associated with the plasma membrane, nucleus and cytoplasm, respectively. Overall, a brief incubation with HU induces a rapid enrichment of α-syn in the nucleus in a subset of cells.

α-Syn increases histone acetylation and RNR levels in HU-treated cells

The involvement of histone acetylation in DNA repair and genome integrity has been shown by work in human cells and yeast (40,41). To test whether α-syn alters histone acetylation upon treatment of cells with HU, western blotting with an antibody specific for histone H3 acetylated at residue K9 was used to assess cells incubated with or without HU (Fig. 4A). To verify that HU induces histone acetylation, H3 acetylation was probed in vector control cells. The acetyl H3 band was significantly more intense in the lysate from −α-syn/+HU cells than from −α-syn/−HU cells (lanes 1 and 2), indicating that HU increases histone acetylation, as expected. H3 acetylation was also probed in ±α-syn cells incubated with HU. The acetyl H3 band was ∼2-fold more intense in the lysate from +α-syn/+HU cells than −α-syn/+HU cells (lanes 2 and 3). α-Syn did not alter histone acetylation in the absence of HU (lanes 4 and 5). The results show that α-syn acts synergistically with HU to increase histone acetylation (Fig. 4B).

Figure 4.

α-Syn increases histone H3 acetylation. (A) α-Syn increases histone acetylation. Cells transformed with the LoTox plasmids were induced for 5 h; the last 2 h with or without HU (200 mm). Lysates were subjected to SDS–PAGE and western blotting. Strain, ɛ1278b (A–D). (B) Plot of α-syn effect on histone acetylation. Quantification and statistical analysis of blots from (A) were conducted by Image J and Excel software, respectively. The (acetyl H3/PGK) band ratio was normalized to 1.0 for the -HU/-α-syn samples. Bars are the average ± s.e.m. of three independent experiments. *P = 0.01; **P = 0.004, two-tailed Student's t-test. (C) α-Syn mutants fail to affect histone acetylation. Cells transformed with the LoTox plasmids were prepared and blotted as in (A). PGK and histone H3 were used as loading controls. (D) Plot of α-syn mutant effects on histone acetylation. Data analysis performed as in (B). Vector control was normalized to 1.0. *P ≤ 0.04, two-tailed Student's t-test.

We also determined whether the α-syn mutants alter histone acetylation. A representative western blot shows the α-syn mutants (Fig. 4C), and an analysis of 3–4 blots facilitated comparison of the mutants (Fig. 4D). HU-treated cells expressing α-syn-▵N, α-syn-▵C or α-syn-NES were very similar to vector control cells in the level of histone acetylation. In contrast, HU-treated cells expressing full-length α-syn exhibited a 2-fold increase in histone acetylation compared with control cells. The results reveal several important points: first, α-syn-NES, which is mostly excluded from the nucleus, fails to alter histone acetylation; second, α-syn-ΔN and α-syn-ΔC can enter the nucleus (see above and Fig. 2A) but only have a slight affect on histone acetylation. The results show that full-length α-syn in the nucleus increases histone acetylation.

We asked whether α-syn can alter the level of RNR proteins in HU-treated cells. To address this issue, a western blot analysis of cells expressing TAP-tagged variants of RNR (Rnr1, Rnr3 and Rnr4) with or without α-syn expression was performed. α-Syn increased Rnr1, Rnr3 and Rnr4 levels 2- to 3-fold in HU-treated cells (Fig. 5A) compared with vector control cells, whereas α-syn failed to increase the level of these three proteins in the absence of HU. Increased RNR levels should counteract the toxic effects of HU.

Figure 5.

α-Syn increases RNR level. (A) Western blot of RNR. TAP-tagged RNR strains transformed with the LoTox α-syn plasmid were induced for 3–4 h with or without HU (200 mm). RNR proteins were visualized with an anti-TAP antibody. Band intensities from 3 to 4 independent experiments, determined using Image J software, were used to calculate the ratio RNR-TAP/PGK. Values in the plot are the average ± s.e.m. of 3–4 independent experiments (*P< 0.05; two-tailed Student's t-test). Growth assay of checkpoint mutants expressing α-syn. W303 deletion strains rad53▵sml1▵ (B), mec1▵sml1▵ (C) and mec1-21 (D) transformed with the LoTox plasmids were serially diluted and spotted on SD, SG and SG + HU plates. Plates were incubated at 30°C for 3 days. SD, synthetic dextrose non-inducing media; SG, synthetic galactose-inducing media.

Interplay between the Mec1–Rad53 signaling and α-syn

The Mec1-Rad53 checkpoint pathway mediates the response to DNA damage (42,43). This checkpoint controls the activity of Sml1 and Rfx1, which are inhibitors of RNR (44). Taking Sml1 as an example, RNR–Sml1 complexes are inactive. When this checkpoint is activated, Sml1 is phosphorylated and rapidly degraded (45), which consequently relieves the inhibition of RNR. To determine whether α-syn functions through this checkpoint, several key deletion strains were tested. Note that the mec1Δ and rad53Δ mutants are lethal, whereas the mec1Δsml1Δ and rad53Δsml1Δ double mutants, and the mec1-21 mutant, are viable (46). With HU, the rad53Δsml1Δ mutant expressing α-syn showed almost identical growth as the vector control cells (Fig. 5B). Similarly, with HU, the mec1▵sml1▵ and mec1-21 cells expressing α-syn showed severe growth defects compared with vector control cells (Fig. 5C and D). The failure of α-syn to protect against HU in these mutants indicates the requirement of Mec1–Rad53 checkpoint and RNR regulation is critical for this protection.

α-Syn maintains Mcm2-RFP foci after HU treatment

The evolutionarily conserved MCM proteins are critical in maintaining genome integrity and cell viability under HU stress by activating dormant or backup origins (47–50). Mcm2–7 proteins form heterohexameric rings that encircle chromosomal replication origins and help recruit the pre-replicative machinery.

To test whether α-syn affects Mcm2–7 complex formation after prolonged exposure to HU, fluorescence microscopy was used to visualize Mcm2-RFP foci in HU-treated, ±α-syn cells. Replication foci and rDNA are marked by CFP-Pol30 and Fob1-YFP, respectively. The number and intensity of the Mcm2-RFP-foci, but not CFP-Pol30 or Fob1-YFP foci, were affected by α-syn. After 24 h incubation in HU, cells expressing α-syn exhibited a significantly larger number of Mcm2-RFP foci and the foci stained more intensely than those in identically treated vector control cells (Fig. 6A–D). No differences in foci were detected between α-syn-expressing cells and vector control cells without HU treatment or cells after a shorter incubation (2 h) in HU (data not shown). The results show that α-syn maintains Mcm2–7 complexes on chromatin after HU treatment, and this could help cells survive HU stress.

Figure 6.

α-Syn maintains Mcm2-RFP foci after HU treatment. α-Syn promotes Mcm2 complex formation. The CFP-Pol30 Fob1-YFP Mcm2-RFP strain transformed with the empty vector (A) or α-syn (B) was grown for 24 h in inducing media supplemented with HU (200 mm). Fluorescence microscopy was used to visualize the different tagged proteins. Plasmids: Lotox (pESC). Scale bar = 5 µm. Arrows show Mcm2 foci. (C) Quantification of Mcm2-RFP foci. Approximately 26% of HU-treated cells expressing α-syn exhibited Mcm2-RFP foci, whereas only 11% of HU-treated vector control cells exhibited Mcm2-RFP foci. Cells (n = 50–100) were counted from two independent experiments. **P< 0.01, two-tailed Student's t-test. (D) Western blot shows α-syn expression. Cells were cultured as described in (A).

Using the growth assay, we tested mutants involved in histone acetylation (rtt109Δ and vps75Δ) (51,52), in the transcriptional regulation of RNR (mbp1Δ and not4Δ) (53,54) and in regulating replication in concert with MCM proteins after HU treatment (not4▵ and dia2Δ) (55,56). That α-syn fails to protect against HU in these mutants (Supplementary Material, Fig. S2) further establishes the importance of histone acetylation, transcriptional regulation of RNR and MCM proteins in enabling α-syn to protect against HU.

α-Syn inhibits HU-induced ROS accumulation

When cells are incubated with HU, ROS accumulates and causes cell death (57–59). Given that α-syn protects cells from the growth inhibitory effects of HU, we asked whether α-syn also inhibits ROS accumulation and used two independent methods to answer this question. Cells were grown for 24 h in inducing media (with or without HU), stained with the ROS sensitive dye DCFH-DA and cells were then examined for ROS using a fluorescence microscope. Without HU, negligible ROS accumulated in cells expressing the LoTox α-syns (vector, α-syn, α-syn-ΔN, α-syn-ΔC or α-syn-NES) (Fig. 7A and B). In contrast, with HU (200 mm), 75% of cells expressing the mutants (ΔN, ΔC or NES) stained for ROS versus only 26% of cells expressing α-syn (Fig. 7A and B). In this assay, α-syn caused a 65% decrease in ROS.

Figure 7.

α-Syn inhibits HU-induced ROS accumulation. (A) ROS measured by fluorescence microscopy. Cells with the indicated LoTox α-syn plasmids were induced for 24 h on plates (galactose), with or without HU (200 mm), and then cells were resuspended in phosphate buffered saline and stained with DCFH-DA. Cells were visualized by differential interference contrast (DIC) and fluorescence microscopy (DCF). Scale bar = 5 μm. (B) Plot of ROS accumulation. Values are mean ± s.e.m of three independent experiments, (n= 600–1000 cells counted; **P< 0.01 versus vector; Student's t-test). (C) ROS measured by flow cytometry. Cells with the indicated LoTox α-syn plasmids were induced for 48 h on plate (galactose), with or without HU (200 mm), and then cells were resuspended in phosphate buffered saline and stained with DCFH-DA. Flow cytometry was used to measure DCF fluorescence triggered by ROS. The y-axis shows the number of cells measured at a given point on the x-axis, which is a logarithmic scale of fluorescence units. (D) HU induces ROS accumulation. The y-axis is the DCF fluorescence signal (normalized to vector control, −HU). Values are the average ± s.e.m. of two independent experiments (*P < 0.05; two-tailed Student's t-test). (E) α-Syn inhibits HU-induced ROS. The y-axis is the DCF fluorescence signal (normalized to vector control, +HU). Values are the average ± s.e.m. of four independent experiments (**P < 0.01; two-tailed Student's t-test). α-Syn-▵N, α-Syn-▵C and α-Syn-NES are labeled as ▵N, ▵C and NES, respectively, in (C) and (E).

To verify these results, identically treated samples incubated with DCFH-DA were also analyzed by flow cytometry, which enables much larger numbers of cells to be probed. Histograms of the various samples with or without HU are shown in Figure 7C, and plots of the normalized mean fluorescence are shown in Figure 2D and E. Without HU, cells expressing the various α-syns (α-syn, α-syn-ΔN, α-syn-ΔC or α-syn-NES) exhibited negligible ROS accumulation compared with vector control cells (Fig. 7C and D). In contrast, with HU (200 mm), as judged by the histograms, vector control cells and the mutants (ΔN, ΔC, and NES) exhibited intense ROS accumulation, whereas identically treated cells expressing α-syn exhibited much less (Fig. 7C). A quantitative analysis of the histograms shows that for cells treated with HU, α-syn caused a significant decrease (50%) in ROS, whereas the mutants (ΔN, ΔC and NES) did not appreciably decrease ROS (Fig. 7E). The two very different methods show reproducibly lower levels of ROS in cells expressing α-syn, which indicates that α-syn inhibits HU-induced ROS accumulation.

DISCUSSION

α-Syn protects cells from HU

The results from this study show that α-syn translocates into the nucleus in response to HU and carries out at least two functions. First, α-syn increases histone acetylation and RNR levels (Figs 4 and 5). Histone acetylation is the key molecular event linked to α-syn-mediated protection against HU, because when either the lysine acetyltransferase RTT109 or the histone chaperone VPS75 are deleted, protection is lost (Supplementary Material, Fig. S2). Although we did not directly detect α-syn-histone complexes, several studies have shown that α-syn binds to histones (37,39). One possibility is that α-syn binds to free histones, which in turn decreases the pool of histones available for binding to DNA. This could destabilize nucleosomes and make them more susceptible to acetylation, resulting in the transcription of certain protective genes, such as RNR. Increased RNR levels counteract HU and promote growth and protect cells from ROS accumulation (Fig. 7). Second, α-syn facilitates the binding of Mcm2 (and Mcm3–7) proteins to chromatin (Fig. 6). This can trigger the firing of dormant origins of replication, which can help cells survive from HU replication stress. One possibility is that α-syn, which has chaperone activities (13), directly binds to Mcm2–7 proteins and stabilizes their association with chromatin. Otherwise, α-syn could up-regulate the expression of Mcm proteins, and more Mcm2–7 complexes bind to chromatin by a mass action effect (discussed subsequently). Future experiments employing mass spectroscopy might determine if α-syn binds to MCM proteins in response to HU. α-Syn influences the highly conserved DNA damage response to protect cells from HU.

A recent study used gene expression profiling to identify genes affected by knocking down the α-syn protein in a human neuroblastoma cell line (SH-SY5Y) (60). Eighty-two up-regulated and 279 down-regulated transcripts were identified. The authors compared their results with results from a whole brain expression profile of α-syn knockout mice and found five overlapping genes. Two of these genes are relevant to this study. In mice and human cells, knocking down the α-syn protein causes a 1.87- and 1.75-fold decrease in histone 1 H3f and Mcm6 transcripts, respectively. Histone 1 H3f is involved in nucleosome assembly and chromosome organization, and α-syn binds to this histone; whereas, Mcm6 is involved in DNA replication and transcription. The implication of this study is that if α-syn aggregates or enters Lewy bodies, the level of the soluble monomer goes to zero, and consequently the histone 1 H3f and Mcm6 transcripts and protein levels are down-regulated. Strikingly, this study showed that decreasing the level of α-syn decreases genome stability. Such results raise the possibility that increasing the level of α-syn increases genome stability (by up-regulating histone 1 H3f and Mcm6). Indeed, our results show that α-syn promotes genome stability by protecting cells from the genotoxin HU.

Nuclear localization of α-syn

Paraquat is an herbicide and mitochondrial poison that has been linked to sporadic PD (61,62). Using midbrain sections of mice, which had been infused with paraquat or vehicle, Goers et al. (39) demonstrated that paraquat triggered up-regulated expression of α-syn, nuclear localization of α-syn and α-syn co-localization with acetylated H3 histones. It was concluded that α-syn could ‘lead to increased transcription and production of proteins in response to a variety of stimuli, including toxic insults.’ The same study demonstrated that α-syn forms a tight 2:1 complex with purified histones in vitro, and that histones dramatically accelerate the rate of α-syn fibrillization. Although no α-syn fibers were detected in cell nuclei in response to paraquat treatment, it was suggested that α-syn fiber formation could severely damage nuclei and underlie the ability of pesticides and herbicides to promote α-syn toxicity. The Goers study was the first to reveal that α-syn has the potential to be protective or toxic in the nucleus. We propose that the concentration of α-syn controls the balance between protection and toxicity when cells are exposed to toxins such as paraquat or HU. These toxins trigger translocation of α-syn into the nucleus, and low levels of α-syn can protect cells by promoting the transcription of certain genes, whereas high levels of α-syn can harm cells by promoting the formation of toxic α-syn-histone nuclear aggregates.

Kontopoulos et al. (37) reported that α-syn inhibits histone acetylation and causes neurotoxicity in cell culture and Drosophila models. It is useful to identify the differences between our study and the Kontopoulos study that explain how α-syn, depending on the conditions, can inhibit or induce histone acetylation. First, the level of α-syn expression was probably much different in the two studies. We surmise that the level of α-syn was much lower in our study than in the Kontopoulos study. Because the high toxicity of the HiTox constructs (Fig. 1A–C) counteracts the protective effect of α-syn after HU treatment, we avoided using the HiTox constructs in most cases; instead, we used the LoTox constructs which are not toxic to yeast. Second, a major difference is that our α-syn-expressing yeast cells were treated with HU, whereas cells and transgenic flies in the Kontopoulos study were not. LoTox α-syn failed to increase histone acetylation in the absence of HU (Fig. 4A and B), suggesting that α-syn functions synergistically with HU to increase histone acetylation. Our view is that the ability of α-syn to alter histone acetylation depends on the type of cell, the concentration of α-syn and the presence of specific chemical stressors. By altering the proteome of a cell or organism, chemicals like paraquat and HU may shift α-syn from inhibiting histone acetylation to promoting it.

This study evaluated the ability of LoTox α-syn mutants, which are non-toxic to yeast cells, to protect against HU. Several points can be made about these mutants. First, α-syn mutants A30P, A53T and S129A/D protect against HU to the same extent as the wild-type protein (Fig. 1D). That S129A or S129D did not affect protection against HU indicates that the phosphorylation status of S129 is not important for protection. Second, compared with full-length α-syn, α-syn-ΔC and especially α-syn-NES are mostly excluded from the nucleus (Fig. 2B). Neither of these constructs protected against HU in any of our assays, indicating that full-length α-syn is required in the nucleus to protect against HU. Third, low levels of α-syn pathogenic mutants (A30P and A53T), which are non-toxic to yeast, have the ability to protect cells from HU (Figs 1 and 2). Strikingly, the key to protection against HU is an intact N-terminus, i.e. A30P and α-syn protect against HU in the growth assay, whereas neither protect when residues 2–10 are deleted (Fig. 2C). Because the α-syn-ΔN construct translocates into the nucleus on HU treatment (Fig. 2A and B), loss of the protection against HU is directly linked to the loss of the N-terminal residues. These results indicate that nuclear localization and full-length structure are required for the protective function of α-syn against HU.

α-Syn can trigger or inhibit ROS accumulation

It is well known that α-syn can trigger ROS accumulation (23,24,63). This is the first study to show that α-syn can also inhibit ROS accumulation induced by a toxin. For example, elevated levels of α-syn induce ROS accumulation, inhibit growth and kill cells (Figs 1 and 2), and deletion of residues 2–10 abolishes this effect. In contrast, low levels of α-syn (and even various mutants) inhibit HU-dependent ROS accumulation and promote growth, and deletion of residues 2–10 abolishes this effect (Figs 2, 4 and 7). Strikingly, toxicity and protection each requires α-syn to have an intact N-terminus. Given that residues 2–10 catalyze helix formation in the N-terminus of α-syn (10), we propose that both toxicity and protection (against HU) are intimately connected to the ability of α-syn to adopt an α-helical conformation. The α-helical conformation of α-syn is required for binding to negatively charged phospholipids in micelles and bilayers (64); perhaps the same conformation is required for binding to a target protein in the nucleus. α-Syn binding to a membrane can be toxic (by poking holes in the membrane) (26,65), whereas α-syn binding to a target protein in the nucleus, such as histones, can be protective (by inducing gene transcription).

The pro-death/pro-survival activities of α-syn could be due to a high affinity Cu²⁺ binding site located at residues 2–9 (66). A recent report suggested that Cu²⁺ binding at residues 2–9 promotes α-syn binding to membranes (67), and another report showed that α-syn, because of its ability to bind metals, is a ferrireductase (33). Since toxicity and protection (against HU) are lost when the high affinity Cu²⁺ binding site (residues 2–10) is deleted, perhaps the bound Cu²⁺ controls toxicity and protection.

In sum, the results reported here may have relevance to both PD and cancer. Regarding PD, one idea is that sporadic PD comes about because α-syn binds to and pokes holes in membranes. That HU can drive α-syn off the plasma membrane in yeast cells raises the possibility that HU might do the same in human cells and thereby lessen α-syn toxicity. Regarding cancer, HU is used for cancers of the blood as well as sickle-cell disease and AIDS (68,69), and α-syn is expressed in blood cells and several kinds of cancers. That α-syn alters the response to HU in yeast, raises the possibility that α-syn also alters the response to HU in human cells. Perhaps synucleins (α, β and γ) make cancer cells resistant to HU.

MATERIALS AND METHODS

Yeast strains, media and reagents

The yeast strains used in this study are isogenic or congenic to ɛ1278b (Matα ura3-52 his3), BY4741 (MATa his3▵1 leu2▵0 met15▵0 ura3▵0), W303 (Matα leu2-3,112 ade2-1 can1-100 his3-11,15 ura3-1 trp1-1 RAD5) and KT1112 (Matα leu2 his3 ura3-52) (Table 1). TAP-tagged RNR strains were purchased from Open Biosystems. S. cerevisiae strains were transformed with pESC-GAL or pAG426-GAL plasmids bearing α-syn or various mutants using the lithium acetate method (24). Yeast-nitrogen-based media and dropout mixtures were obtained from Sigma and Qbiogene, and media were prepared according to the standard yeast protocols (70). Media containing glucose (2% w/v) and lacking uracil (or uracil and tryptophan) to maintain selection for plasmids are referred to as non-inducing media (SD, synthetic dextrose media). Inducing media were the same except that galactose (2%) replaced glucose (SG, synthetic galactose media). Restriction enzymes, Pfu DNA ligase and other reagents for DNA manipulation were purchased from Promega and Stratagene. HU was purchased from Sigma or US Biologicals. All other reagents were purchased from Sigma. Cells in liquid media were grown in glass or plastic tubes with shaking at 30°C.

Plasmids

All DNA manipulations followed standard protocols (71). pESC-URA-GAL (Stratagene, # 217454) is a 2 μ plasmid that contains GAL1 and GAL10 yeast promoters; each insert was subcloned behind the GAL1 promoter. The N- and C-terminal truncations of α-syn (α-syn-▵N, α-syn-▵C) were generated by the polymerase chain reaction using Pfu ultra polymerase and subcloned into pESC-URA-GAL at the SalI sites and the BamHI/SalI sites, respectively. A30P, A53T, S129A, S129D and A30P-▵N were generated via site-directed mutagenesis as previously described (16,72) and subcloned into pESC-URA-GAL at the SalI sites. α-Syn-NES was constructed according to the method previously described (73). The pESC-URA-GAL α-syn plasmids are referred to as LoTox plasmids.

pAG426-GAL (Addgene) is a 2 μ plasmid that contains the GAL1 yeast promoter. For pAG426-GAL plasmids, recombination-based Gateway cloning system (Invitrogen, Carlsbad, CA) was used. The coding sequence of α-syn and mutants was amplified by PCR with forward and reverse primers, which consist of the 25-nucleotide (nt) attB sequence and ∼25–30 nt of gene-specific sequence, and then recombined with counter-selectable ccdB gene of donor vector pDONR221 by BP clonase (Invitrogen). α-Syn variants were transferred into Gateway destination vector, pAG426GAL by LR clonase (Invitrogen). The pAG426-GAL α-syn plasmids are referred to as HiTox plasmids.

We also used pRS316-GAL-GFP-α-syn, which is a centromeric plasmid that harbors GFP-α-syn driven by the GAL1 promoter (74).

Growth assay

Strains transformed with the indicated plasmids were pre-grown in non-inducing (raffinose or glucose) media lacking uracil (and/or tryptophan) at 30°C to OD₆₀₀ = 0.6–1. Cells were then washed and resuspended in water to the same OD₆₀₀, serially diluted (1:10) and spotted (5 μl) onto SD and SG plates with or without HU (or other drugs). Plates were incubated for 3–5 days at 30°C.

Western blotting

Cells were pre-grown in synthetic non-inducing media lacking uracil (and/or histidine) and incubated overnight at 30°C. Cells were then washed, transferred to inducing media lacking uracil (and/or histidine) and incubated at 30°C for 3 h (in some cases, incubated overnight). HU (100–200 mm) was then added and cells were incubated for 2–3 h (in some cases, HU was added at the beginning of induction). Cells were harvested, lysed with glass beads and the lysate was subjected to SDS–PAGE and western blotting, as described previously (24). Commercially prepared 4–20% gradient or 15% polyacrylamide gels were purchased from Bio-Rad. Proteins were visualized using a polyclonal anti α-syn antibody (#2642, Cell Signaling), a polyclonal TAP antibody (#CAB1001, Open Biosystems), a polyclonal anti acetyl-histone H3 (Lys9) antibody (#9671, Cell Signaling) and a monoclonal anti-3-phosphoglycerate kinase (PGK) antibody (#459250, Invitrogen), for the loading control. Proteins were visualized using the Amersham™ ECL™ detection kit (RPN2106, GE Healthcare) in conjunction with autoradiography film and X-ray film processor (Konica SRX-101A). The Image J program was used for signal quantification. P-values were determined using the two-tailed Student's t-test. Representative blots from several experiments are shown in the figures.

Fluorescence microscopy

Fluorescence microscopy was performed with an Olympus AX70 microscope, equipped with an Olympus UPlanFI 100×/1.35 NA objective and a Roper CoolSNAP HQ CCD camera. A 41001 filter (Chroma Technology) was used for imaging GFP, a U-MNG filter (Olympus) was used for RFP (mCherry) and a JP4 filter (Chroma Technology) was used for CFP and YFP. The acquisition software was Slidebook 4.0 (Intelligent Imaging Innovations).

ROS accumulation assay

ɛ1278b strains transformed with the various plasmids were grown on solid plates containing inducing media (galactose) lacking uracil and supplemented with 200 mm HU for 24 h at 30°C. Cells were scraped off the plates, washed twice in distilled water and then incubated in distilled water containing 10 μg/ml DCFH-DA for 1 h at room temperature. DCFH-DA, which is cell permeable and non-fluorescent, is oxidized by intracellular ROS to the fluorescent compound 2′,7′-dichlorofluorescein (DCF) (75). Cells were washed and visualized by differential interference contrast (DIC) and fluorescence microscopy (DCF). Similarly treated cells were analyzed by flow cytometry, with minor modifications of the protocol. Cells were scraped off the plates after 48 h and resuspended in 1 ml phosphate-buffered saline (PBS) and incubated with 10 µg/ml DCFH-DA for 1 h at room temperature. Cells were washed and resuspended in PBS prior to the analysis. The ROS-induced DCF fluorescence intensity of 2 to 5 × 10⁴ cells was measured by flow cytometry (Becton-Dickinson FACSCalibur), as described previously (76). Data acquisition was performed using CellQest Pro software, and FlowJo (v 9.3.2) software was used to analyze the histograms. The experiments were repeated four times.

TEM and immunogold labeling

Fixation, immunogold labeling and TEM was performed as previously described (74). The primary anti-α-syn antibody (Cell Signaling Technology) and secondary 10 nm gold-conjugated goat anti-rabbit IgG (Jackson Immuno-Research Laboratories) were used for immunogold labeling. A total of 30 yeast cells, from two independent experiments, were analyzed for gold particle distribution. Three areas from the membrane, cytoplasm and nucleus were randomly chosen. Gold particles were manually counted using the NIH ImageJ software (http://rsb.info.nih.gov/ij/) and the particle density was calculated as the number of gold particles divided by the selected area (particles/um²). Based on the estimated size of a 10 nm gold particle and the epitope (∼30 nm) (77), the area of the membrane was determined as 60 nm for particle measurement, i.e. 30 nm from each side of the membrane. The percentage of α-syn-gold particles in the nucleus is given by [particles in the nucleus / (particles in the nucleus + particles in the cytoplasm + particles on the plasma membrane)] × 100.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported by the National Institutes of Health (NS057656 to S.N.W.) and by the Parkinson's Resource of Northwest Louisiana.