Tyrosine and serine phosphorylation of α-synuclein have opposing effects on neurotoxicity and soluble oligomer formation (original) (raw)
α-Synuclein is phosphorylated at Tyr125. We used an antibody that specifically recognizes α-synuclein phosphorylated at Tyr125 (PY125 antibody) to demonstrate tyrosine phosphorylation in transgenic Drosophila expressing wild-type human α-synuclein (α-synWT; Figure 1A). We next created transgenic Drosophila lines carrying mutant α-synuclein complementary DNA constructs in which the COOH-terminal codons for Tyr125 alone (α-synY125F) or all 3 tyrosine residues (Y125, Y133, and Y136) had been replaced by codon for phenylalanine to prevent phosphorylation (α-synYF) (18). Specificity of the antibody was supported by lack of immunoreactivity following substituting Tyr125 to phenylalanine (Figure 1A). In addition, pretreatment of the homogenate from the brains of flies expressing wild-type α-synuclein with phosphatase to remove phosphate groups also abolished immunoreactivity (Figure 1B). These findings taken together strongly support the specificity of the PY125 antibody in these studies.
Blocking tyrosine phosphorylation increases α-synuclein toxicity. (A) The PY125 antibody specifically recognizes phosphorylated Tyr125. Phenylalanine substitution at Tyr125 eliminates immunoreactivity. Lane 1, α-synWT; lane 2, α-synY125F. The same membrane was stripped and reprobed with a phosphorylation-independent α-synuclein antibody (anti-syn). (B) The α-synYF mutant effectively eliminates anti-PY125 immunoreactivity. Compare lane 1 (α-synWT) with lane 3 (α-synY125F). Lane 2 (α-synWT) is the same as lane 1 except for pretreatment with phosphatase. Flies were 20 days old. (C) Quantitative analysis of TH-immunoreactive dorsomedial dopamine neurons over time in transgenic flies. Values represent mean ± SEM. Accelerated loss of TH-immunoreactive neurons is observed in 10-day-old α-synYF transgenic flies (*P < 0.05, multivariant ANOVA with supplementary Newman-Keuls test). The driver in A–C was elav-GAL4. (D and E) Blocking tyrosine phosphorylation enhances retinal degeneration. Tangential section through the retina of an aged adult fly expressing α-synWT shows mild architectural distortion (D), whereas expression of α-synYF produced a greater loss of retinal integrity with large vacuoles (E). Original magnification in D and E, ×400. (F) Quantitative comparison demonstrating significant reduction in the percentage of normal photoreceptor clusters in α-synYF transgenic flies (*P < 0.01, multivariant ANOVA with supplementary Newman-Keuls test). Flies were 30 days old. The driver in D–F was GMR-GAL4. (G) Accelerated loss of climbing ability in transgenic flies expressing α-synS129D and α-synYF throughout the nervous system (elav-GAL4 driver). Asterisks indicate climbing scores that are significantly different from the control and wild-type α-synuclein transgenic scores (*P < 0.01, multivariant ANOVA with supplementary Newman-Keuls test). Control genotype: elav-GAL4/+. Error bars in F and G represent SEM.
Tyrosine phosphorylation protects from α-synuclein neurotoxicity. To determine the role of tyrosine phosphorylation in α-synuclein neurotoxicity and aggregation, we used the binary GAL4/UAS system to direct tissue-specific transgene expression (19). Lines expressing equivalent levels of α-synWT and α-synYF were used in subsequent studies (Figure 1B and Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI39088DS1). First, the elav-GAL4 driver was used to target expression in a pan-neural pattern. Brains from transgenic flies 1, 5, 10, 15, 20, and 30 days after eclosion were then immunostained with an antibody against tyrosine hydroxylase (TH), which specifically identifies dopaminergic neurons. We focused our analysis on the dorsomedial cluster of dopaminergic neurons, because these neurons are readily identifiable and are preferentially sensitive to α-synuclein toxicity (15, 20, 21). In 1-day-old and 5-day-old transgenic flies expressing α-synYF, the number of dopaminergic neurons in the dorsomedial clusters was not significantly different from the number present in young flies expressing α-synWT or in nontransgenic controls (Figure 1C). Thus, expression of tyrosine phosphorylation–incompetent α-synuclein does not disrupt development of dopaminergic neurons. However, onset of neurodegeneration was significantly accelerated. On average, 2 TH-immunoreactive neurons were observed in 10-day-old flies expressing α-synYF, whereas dorsomedial dopaminergic neurons were relatively well preserved in wild-type α-synuclein transgenic lines at the same time point (Figure 1C). Similar to flies expressing α-synWT, those expressing α-synYF exhibited a marked loss of TH-immunoreactive neurons by 20 days of age (Figure 1C). No further loss was noted at 30 days (data not shown).
With the pan-neural elav-GAL4 driver we have not previously seen significant toxicity of α-synuclein in brain cells other than dopaminergic neurons (15), even when we have analyzed a mutant form of α-synuclein with increased neurotoxicity or genetic modifiers that enhance dopaminergic cell death (14, 22). To determine whether the enhanced toxicity of α-synYF retained specificity for dopaminergic neurons, we prepared serial sections through the whole brain and stained the sections with hematoxylin to identify neuronal and glial cell bodies in the cortex and eosin to highlight neuropil structures. These studies revealed normal overall development of the brain, with appropriate size and configuration of major brain structures. In addition, there was no clear loss of nuclei in the cortex, sign of DNA fragmentation, or loss of neuropil integrity (data not shown). Thus, enhanced α-synuclein toxicity to dopaminergic neurons was not accompanied by widespread, general neurodegeneration in the brain.
Expression of α-synWT targeted to the retina using a glass gene promoter construct (GMR-GAL4 driver) caused mild adult-onset retinal degeneration manifested by disruption of the normal arrangement of ommatidia (Figure 1D) (15). As would be predicted based on the increased toxicity seen to dopamine neurons, expression of α-synYF produced a more prominent retinal degeneration, with vacuole formation and marked loss of tissue integrity (Figure 1E). To evaluate retinal degeneration on a quantitative basis, we determined the percentage of normal ommatidia, which was significantly decreased in flies expressing α-synYF, compared with α-synWT transgenics or control flies not expressing α-synuclein (Figure 1F).
As an additional means to assess the toxicity of α-synYF, we evaluated the locomotor performance of flies expressing the mutant variant. Transgenic flies that express either the wild-type or familial Parkinson disease–linked mutant form of α-synuclein in their central nervous systems, via the pan-neural elav-GAL driver (15), or specific dopaminergic drivers (23, 24), show an age-dependent reduction in climbing ability when compared with controls. We found that flies expressing α-synYF showed a greater reduction in climbing ability compared with flies that expressed wild-type α-synuclein (Figure 1G), demonstrating the ability of α-synYF to accelerate functional decline as well as neuropathology.
α-Synuclein can be phosphorylated at C-terminal tyrosines by a number of protein tyrosine kinases, including Syk (18) and Src family members (16). To explore the role of kinases in controlling α-synuclein toxicity in vivo and to confirm that the enhanced toxicity of α-synYF was mediated through reduced tyrosine phosphorylation, we used the GAL4/UAS system to overexpress shark, a Drosophila homolog of Syk (25). Overexpression of shark increased phosphorylation at Tyr125 on α-synWT (Figure 2A). In addition, overexpression of shark was also able to elevate Tyr125 phosphorylation of a mutant version of α-synuclein that mimics phosphorylation at Ser129 (α-synS129D). Levels of total α-synuclein were unchanged by the expression of shark (Figure 2A). Coexpression of shark significantly rescued the neurotoxicity of both α-synWT and α-synS129D (Figure 2, B and C). Thus, increasing the expression of Src tyrosine kinase increased phosphorylation of α-synuclein in vivo and ameliorated its selective neurotoxicity.
Overexpression of Drosophila tyrosine kinase shark increases α-synuclein phosphorylation and ameliorates its neurotoxicity. (A) Increased tyrosine phosphorylation in flies coexpressing shark and α-synuclein in the brain compared with flies expressing α-synuclein alone. Flies were 1 day old. (B–D) Quantitative analysis of TH-immunoreactive neurons in flies coexpressing shark and α-synuclein and flies expressing α-synuclein alone. Values represent mean ± SEM. Asterisks indicate that the difference in TH-immunoreactive dopaminergic neuronal number between the 2 genotypes at the time points examined is statistically significant (*P < 0.01, **P < 0.05, multivariant ANOVA with supplementary Newman-Keuls test). The driver in A–D was elav-GAL4. (E–G) Augmented tyrosine phosphorylation suppresses retinal toxicity. Retinal degeneration in flies expressing α-synS129D (E) is largely prevented by coexpressing shark (F). Original magnification in E and F, ×400. (G) Quantitative comparison demonstrating significant rescue of the percentage of normal photoreceptor clusters in α-synS129D transgenic flies expressing shark (*P < 0.01, Student’s t test). Error bars represent SEM. The driver in E–G was GMR-GAL4. Flies were 30 days old.
In contrast, no significant rescue was observed upon coexpression of shark and α-synYF (Figure 2D), indicating that the rescue was indeed mediated via tyrosine residues and supporting the specificity of shark action. Similarly, targeting shark expression to the eye decreased the retinal toxicity of α-synuclein. Retinal cell loss was increased in flies expressing α-synS129D compared with α-synWT (Figure 2E). Retinal degeneration was attenuated by coexpression of shark, as demonstrated by representative histological sections (Figure 2F) and quantitative analysis (Figure 2G).
Opposing effects on soluble oligomer formation. We have recently identified soluble oligomeric species of α-synuclein in transgenic Drosophila (22). Formation of these oligomer species requires the central NAC aggregation domain of α-synuclein (26). To determine the role that α-synuclein phosphorylation plays in controlling oligomer formation in vivo, we first characterized the levels of oligomeric species in flies expressing α-synS129D and the Ser129 phosphorylation–incompetent mutant α-synS129A, which is not toxic in our system (14). Western blot analysis of high-speed supernatants of fly brain homogenates did not identify oligomeric forms in 20-day-old nontransgenic controls, or in 1-day-old α-synuclein transgenic flies. In contrast, oligomeric species appeared in 10-day-old α-synuclein transgenic flies and accumulated with age. Levels of soluble oligomers correlated well with α-synuclein neurotoxicity. Soluble oligomers accumulated rapidly in flies expressing α-synS129D and were significantly less abundant in flies expressing α-synS129A (Figure 3, A and B). Preventing Ser129 phosphorylation with the α-synS129A mutant did not completely block oligomer formation, supporting the ability of α-synS129A to have toxicity under other circumstances (see Discussion). Coexpression of wild-type α-synuclein and the serine kinase Gprk2 produced an enhancement of α-synuclein oligomerization similar to that seen with expression of α-synS129D (Figure 3C), consistent with a role for endogenous phosphorylation at Ser129 in controlling oligomerization.
α-Synuclein toxicity correlates with accumulation of soluble high-molecular-weight oligomers. (A) Representative immunoblot of soluble fractions from brain homogenates run on a 4%–12% gradient gel shows increased oligomer formation in flies expressing α-synS129D and decreased oligomer formation in flies expressing α-synS129A. Aliquots of the same protein preparations were loaded on a separate 10%–20% gel to visualize monomeric α-synuclein and tubulin. Control genotype: elav-GAL4. (B) Quantitative analysis of oligomer formation using densitometry. Values represent mean ± SEM of 3 independent experiments. Single asterisk indicates the significantly increased accumulation of high-molecular-weight species in flies expressing α-synS129D (*P < 0.01, multivariant ANOVA with supplementary Newman-Keuls test) at 10 days. Double asterisks indicate the significantly decreased accumulation in α-synS129A flies at 20 days (**P < 0.01, multivariant ANOVA with supplementary Newman-Keuls test). (C) Coexpression of the serine kinase Gprk2 with α-synWT increases high-molecular-weight α-synuclein species. Flies were 10 days old.
To determine whether Tyr125 phosphorylation also altered α-synuclein oligomerization, we compared transgenic lines expressing α-synWT and α-synYF. Preventing Tyr125 phosphorylation with the α-synYF mutation promoted earlier accumulation of oligomeric species compared with flies expressing the wild-type protein (Figure 4, A and B). Conversely, increasing Tyr125 phosphorylation by expression of shark kinase decreased soluble oligomer formation of both α-synWT and α-synS129D (Figure 4, C and D). Thus, the formation of soluble oligomers was regulated in an antagonistic fashion by phosphorylation of Ser129 and tyrosines in the carboxyterminal domain of α-synuclein. Consistent with a role for phosphorylation of Ser129 in promoting oligomer formation and phosphorylation of Tyr125 in preventing oligomerization, oligomers were preferentially phosphorylated at Ser129 compared with Tyr125 (Supplemental Figure 2).
Tyr125 phosphorylation status alters α-synuclein oligomerization. (A) Increased high-molecular-weight α-synuclein species are evident in flies expressing α-synYF. A lighter exposure than in Figure 3A is shown to demonstrate the increase in oligomeric species in α-synYF. Control genotype: elav-GAL4. (B) Quantitative analysis of oligomer formation. Values represent mean ± SEM of 3 independent experiments. Asterisk indicates significant difference compared with α-synWT (*P < 0.01, multivariant ANOVA with supplementary Newman-Keuls test). (C) Coexpression of the tyrosine kinase shark decreases high-molecular-weight α-synuclein species formation of both α-synWT and α-synS129D. (D) Quantitative analysis of shark kinase inhibition of oligomer formation. Values represent mean ± SEM of 3 independent experiments. Asterisks indicate significant decrease in flies coexpressing shark and α-synuclein (*P < 0.05, Student’s t test). Flies were 10 days old.
To further investigate the role of Tyr125 phosphorylation in modulating the solubility of α-synuclein, we monitored the formation of large insoluble protein aggregates, or inclusion bodies, by immunostaining fly brains with antibodies directed against α-synuclein (15, 22). We saw no clear alteration in the number of large inclusions in α-synYF transgenic flies. Nor was there any detectable increase in proteinase K resistance, a typical feature of fibrillar protein aggregates (27) (data not shown).
To determine whether sufficient tyrosine phosphorylation of α-synuclein is present in our system to support a plausible role for tyrosine phosphorylation in disease pathogenesis, we evaluated the amount of α-synuclein that was phosphorylated at Tyr125 in vivo. We performed 2-dimensional gel electrophoresis followed by Western blotting using fresh fly brain homogenates prepared in the presence of phosphatase inhibitors. We found that approximately 30% of total α-synuclein was phosphorylated at Tyr125 (Supplemental Figure 3), supporting the plausibility of a role for Tyr125 phosphorylation in controlling α-synuclein aggregation and neurotoxicity in vivo. A previous study of human postmortem tissue failed to identify tyrosine-phosphorylated α-synuclein (28). We therefore evaluated the sensitivity of Tyr125 phosphorylation in our system by incubating postmortem tissue in the absence of phosphatase inhibitors for various periods of time. We observed that the reactivity of α-synuclein to anti-PY125 decreased with increasing intervals of incubation, consistent with postmortem dephosphorylation at Tyr125 (Supplemental Figure 4).
Possible mechanisms of neuroprotection. Although we demonstrate that tyrosine phosphorylation decreases oligomer formation, the way in which phosphorylation controls aggregation is not clear. One possibility is that phosphorylation of α-synuclein at Tyr125 might protect from neurotoxicity and oligomer formation by decreasing phosphorylation at Ser129. We thus compared the degree of Ser129 phosphorylation in flies with varying levels of Tyr125 phosphorylation. We detected no clear differences in the levels of phospho-Ser129 among flies expressing α-synWT, α-synYF, and α-synWT coexpressed with shark kinase to increase Tyr125 phosphorylation (Supplemental Figure 5, A and C). We also detected no influence of Ser129 phosphorylation on phosphorylation of Tyr125. Similar levels of phospho-Tyr125 were present in flies expressing α-synWT, α-synS129A, and α-synS129D (Supplemental Figure 5, B and D). The ability of the PY125 antibody to recognize α-synS129A and α-synS129D to the same degree as α-synWT further supports the specificity of the antibody.
Alternatively, phosphorylation of α-synuclein at Tyr125 might influence proteolytic processing of α-synuclein. However, we found no clear alteration in levels of C-terminally truncated α-synuclein in flies expressing α-synYF or flies coexpressing shark kinase along with α-synWT to increase phosphorylation of Tyr125 (Supplemental Figure 6). To further explore the role of phosphorylation in controlling C-terminal proteolysis of α-synuclein, we examined the levels of truncated protein in flies expressing α-synS129A or α-synS129D, as well as flies coexpressing α-synWT and the kinase Gprk2 to increase phosphorylation at Ser129. In these experiments, we found no evidence that modification of Ser129 altered levels of C-terminally truncated α-synuclein (Supplemental Figure 6).
Tyr125 phosphorylation diminishes during human aging. To explore the relevance of Tyr125 phosphorylation of α-synuclein in humans, we examined the levels of phosphorylation in frontal cortical homogenates from young (18–28 years) and aged (69–86 years) normal adult controls. We found a striking reduction in Tyr125 phosphorylation with aging (Figure 5, A and B). Remarkably, we saw a similar aging-related decline in α-synuclein phosphorylation in transgenic flies (Figure 5C). The age-related reduction in tyrosine phosphorylation of α-synuclein appears to be specific, because overall tyrosine phosphorylation was not significantly decreased in aged Drosophila (data not shown). Increasing age is an important risk factor for developing symptomatic α-synuclein neurotoxicity in both humans and flies (15, 29).
α-Synuclein is phosphorylated at Tyr125, and phosphorylation is reduced in aging human and Drosophila brain. (A) Age-related tyrosine phosphorylation levels in humans. Shown are immunoblots from 5 young (≤28 years old) and 5 aged (≥69 years old) frontal cortical samples as detected by the antibody specific for α-synuclein phosphorylated at Tyr125. (B) The densitometry intensities of anti-PY125–positive bands pooled from both young and aged groups. Values represent mean ± SEM of 3 independent experiments (*P ≤ 0.01, Student’s t test). (C) The PY125 blot demonstrates age-related decline in tyrosine-phosphorylated wild-type α-synuclein in Drosophila brain extracts. The driver was elav-GAL4. Days after eclosion are indicated. (D) Tyr125 phosphorylation is further reduced in dementia with Lewy body (DLB) patients. Shown are immunoblots of frontal cortical samples from 5 DLB cases (≥65 years old) and 5 aged controls (≥69 years old).
We next compared the levels of Tyr125 phosphorylation in brain homogenates from normal aged controls and patients with dementia with Lewy bodies, a synucleinopathy with prominent cortical Lewy bodies (30). We did not detect phosphorylation at Tyr125 in frontal cortical homogenates from 5 patients (65–82 years) (Figure 5D), in contrast to clear immunoreactivity in aged controls (69–86 years). Thus, loss of protective tyrosine phosphorylation may predispose to clinically relevant α-synuclein neurotoxicity in human disease.