14-3-3θ phosphorylation exacerbates alpha-synuclein aggregation and toxicity

Abstract

Aggregation of alpha-synuclein (αsyn) plays an integral role in Parkinson’s disease (PD) and Dementia with Lewy bodies (DLB). 14-3-3θ is a highly expressed brain protein with chaperone-like activity that regulates αsyn folding. 14-3-3θ overexpression reduces αsyn aggregation, transmission between cells, and neuronal loss, while 14-3-3 inhibition promotes αsyn pathology. We previously observed increased 14-3-3θ phosphorylation at serine 232 in human PD and DLB brains. Here we examine 14-3-3θ phosphorylation’s effects on αsyn aggregation and toxicity. Using a paracrine αsyn model, we found that the non-phosphorylatable S232A 14-3-3θ protected while the phosphomimetic S232D 14-3-3θ failed to protect against αsyn paracrine toxicity. The S232A mutant reduced oligomerization of released αsyn while the S232D mutant did not. The S232D mutant showed significant reduction in αsyn binding compared to wildtype or S232A 14-3-3θ. Using knock-in mouse models expressing the S232A or S232D mutation in the cortex and hippocampus, we examined the impact of S232 phosphorylation on αsyn aggregation in the αsyn preformed fibril (PFF) model. Primary neurons from S232D mice showed increased αsyn inclusion formation compared to neurons from Cre control mice upon PFF treatment. In contrast, neurons from S232A mice showed reduced αsyn inclusions. αSyn PFF injection into the dorsolateral striatum induced higher αsyn inclusion numbers in the sensorimotor cortex of S232D mice compared to Cre control mice. In conclusion, 14-3-3θ phosphorylation at S232 interrupts the ability of 14-3-3θ to bind and regulate αsyn aggregation. Increased 14-3-3θ phosphorylation observed in human PD and DLB likely accelerates neurodegeneration in these disorders.

1. Introduction

Alpha-synuclein (αsyn) is critical to neurodegeneration in Parkinson’s disease (PD) and Dementia with Lewy bodies (DLB). This protein is the key component of Lewy bodies, the pathological hallmark observed in these disorders (Spillantini et al., 1997). Growing evidence points to its tendency to aggregate and its potential for cell-to-cell transmission as critical to the pathogenic process (Brás and Outeiro, 2021; Choi et al., 2021; Leak et al., 2024; Neupane et al., 2023; Uemura et al., 2020; Wu and Schekman, 2024). However, homeostatic mechanisms that normally prevent αsyn aggregation and transmission from cell to cell are not well understood.

One important family of proteins that colocalize with αsyn in Lewy bodies are the 14-3-3 proteins (Berg et al., 2003; Kawamoto et al., 2002). 14-3-3 s make up a highly conserved family of seven proteins that play important roles in protein folding and protein trafficking, among other functions (Giusto et al., 2021; Mrowiec and Schwappach, 2006; Pair and Yacoubian, 2021; Sluchanko et al., 2014; Vincenz and Dixit, 1996; Yano et al., 2006). The 14-3-3θ isoform reduces recombinant αsyn fibrillization in vitro (Plotegher et al., 2014). We previously demonstrated that 14-3-3θ regulates the cell-to-cell transmission of αsyn in multiple in vitro and in vivo models (Underwood et al., 2021; Wang et al., 2018). Using a novel paracrine model, in which induction of increased αsyn levels results in the release and uptake of oligomeric αsyn that is toxic to separately cultured neurons, we found that, despite higher total amounts of αsyn release, 14-3-3θ overexpression reduced αsyn oligomerization, seeding capability, and toxicity in primary neurons (Wang et al., 2018). Conversely, 14-3-3 inhibition increased αsyn oligomerization, seeding, and neuronal death in this paracrine system (Wang et al., 2018). Similarly, 14-3-3θ overexpression reduced αsyn aggregation, neuronal death, and transmission between cells in primary neuronal cultures treated with recombinant αsyn preformed fibrils (PFFs), while 14-3-3 inhibition promoted αsyn pathology in this model (Wang et al., 2018). In vivo, both adeno-associated viral mediated overexpression and transgenic overexpression of 14-3-3θ slowed αsyn aggregation, rescued behavioral deficits, and reduced neuronal loss in response to striatal αsyn PFF injection (Underwood et al., 2021). In contrast, 14-3-3 inhibition accelerated αsyn aggregation, behavioral deficits, and neuronal loss in vivo (Underwood et al., 2021).

A critical question is how 14-3-3θ may lose its ability to regulate αsyn misfolding in PD and DLB. Examination of 14-3-3 proteins in human PD and DLB temporal cortical brain lysates revealed a dramatic increase in the phosphorylation of 14-3-3θ at serine 232 in Triton X-100 insoluble fractions (McFerrin et al., 2017). This increase in S232 phosphorylation correlated with reduced cognitive function and with increased αsyn pathology (McFerrin et al., 2017). We previously demonstrated that rotenone, a pesticide which increases the risk of PD, induces an initial increase in 14-3-3θ phosphorylation at S232 (Slone et al., 2015). 14-3-3θ overexpression is protective against rotenone toxicity (Slone et al., 2015; Slone et al., 2011; Yacoubian et al., 2010), but the phosphomimetic S232D 14-3-3θ mutant fails to protect against rotenone and other neurotoxins (Slone et al., 2015). We have additionally observed that αsyn overexpression in our paracrine model induces S232 phosphorylation that precedes the onset of αsyn paracrine toxicity (Slone et al., 2015).

Based on these findings, we hypothesized that increased 14-3-3θ phosphorylation at S232 may disrupt the ability of 14-3-3θ to bind and regulate αsyn. With this disruption in function, αsyn can then misfold and aggregate, leading to increased cellular toxicity. To test this, we examined the potential impact of S232 phosphorylation on 14-3-3θ function using several αsyn models: the paracrine αsyn model, the in vitro PFF model, and the in vivo PFF model.

2. Material and methods

2.1. Cell lines

Isyn cell line was made by transduction of SK-N-BE(2)-M17 (M17) male neuroblastoma cells (obtained and authenticated by ATCC, #CRL-2267; RRID:CVCL_0167) with the doxycycline (doxy)-inducible αsyn pSLIK lentivirus, followed by selection for stable transfection with G418 as previously described (Slone et al., 2015). M17 cells express tyrosine hydroxylase and dopamine that is increased upon differentiation (Carvajal-Oliveros et al., 2022). Isyn cells were grown in 1:1 Eagle’s MEM/F12K containing 10 % fetal bovine serum (FBS), 1 % penicillin/streptomycin, and G418 (500 μg/ml) at 37 °C.

To create isyn/14-3-3θ, isyn/S232A, and isyn/S232D lines, isyn cells were transduced with the doxy-inducible wildtype (WT) 14-3-3θ, mutant S232A 14-3-3θ, or mutant S232D 14-3-3θ hygromycin (hygro)-selectable pSLIK lentiviruses. Selection for stable transfection was performed with hygro (100 μg/ml) in addition to G418. The UAB Neuroscience Core Center constructed the WT 14-3-3θ pSLIK construct. WT 14-3-3θ was first cloned into the pEN_TTmcs vector and then recombined with the hygro-selectable pSLIK lentiviral construct (Shin et al., 2006; Wang et al., 2018). The mutant 14-3-3θ pSLIK constructs were similarly constructed using site directed mutagenesis of the S232 site. Cells were grown in 1:1 Eagle’s MEM/F12K containing 10 % FBS, 1 % penicillin/streptomycin, G418 (500 μg/ml), and hygro (100 μg/ml) at 37 °C. Doxy treatment at 10 μg/ml was used to induce expression of αsyn and 14-3-3θ.

SH-SY5Y cells (obtained and authenticated by ATCC [#CRL-2266 RRID:CVCL_0019]) were grown in 1:1 Eagle’s MEM/F12K containing 10 % FBS and 1 % penicillin/streptomycin. Differentiation of SH-SY5Y cells was induced with retinoic acid (10 μM; Sigma-Aldrich, #R2625) for 5–7 days until cell proliferation stopped and prominent neurite extension was observed.

OptiMEM media with 10 % FBS was used to grow and maintain H4 neuroglioma cells (obtained and authenticated by ATCC [#HTB-148 RRID:CVCL_1239]).

2.2. Mice

Experiments using mice were performed in accordance with the guidelines of the National Institute of Health (NIH) and University of Alabama at Birmingham (UAB) Institutional Animal Care and Use Committee (IACUC). The IACUC at UAB approved animal work performed in this study. A conditional knock-in (KI) mouse expressing the 14-3-3θ S232D phosphomimetic mutant was created in concert with Cyagen. Briefly, a targeting vector containing homology arms and the S232D mutation in exon 5 was created, linearized, and then introduced into C57BL/6 embryonic stem cells via electroporation. Sequence of the S232D targeting vector is found in the Supplemental Material (Supp. Fig. 1). After drug selection, PCR screening, and Southern Blot confirmation, selected ES clones were used to generate germline-transmitted founder lines that expressed the S232D mutation. These mice were bred to each other to maintain a colony of conditional KI S232D (Cre −/− S232D +/+) mice. Similarly, a conditional KI mouse expressing the S232A mutant was created in concert with Cyagen. Sequence of the S232A targeting vector is found in the Supplementary Material (Supp. Fig. 1).

To induce S232D expression in cortical and hippocampal regions of the brain for the studies described here, the conditional KI S232D mouse (Cre −/− S232D +/+) was first crossed with the heterozygous Emx1-Cre mouse (Cre +/− S232D −/−) (Gorski et al., 2002). Heterozygous S232D mice (Cre +/− S232D +/−) were then crossed with Cre −/− S232D +/−mice to generate the four groups of mice required for the studies described in this paper: 1) WT mice (Cre −/− S232D −/−); 2) Cre control mice (Cre +/− S232D −/−); 3) heterozygous S232D KI mice (Cre +/− S232D +/−); and 4) homozygous S232D KI mice (Cre +/− S232D +/+).

To induce S232A expression in cortex and hippocampus, the conditional KI S232A mouse (Cre −/− S232A +/+) was first crossed with the heterozygous Emx1-Cre mouse (Cre +/− S232D −/−) (Gorski et al., 2002). Heterozygous S232A mice (Cre +/− S232A +/−) were then crossed with Cre −/− S232D +/− mice to generate experimental mice required for the studies described in this paper: 1) WT mice (Cre −/− S232A −/−); 2) Cre control mice (Cre +/− S232A −/−); and 3) homozygous S232A KI mice (Cre +/− S232D +/+).

2.3. RT-PCR and sequencing

To confirm RNA expression of the S232A or S232D 14-3-3θ mutants in the appropriate brain regions of Emx1-Cre +/− S232A +/+ mice or Emx1-Cre +/− S232D +/+ mice, we extracted RNA, reverse-transcribed RNA into cDNA, performed PCR for 14-3-3θ spanning the mutation site, and then sequenced the PCR product. RNA was extracted from the cortex and hippocampus using the RNeasy Mini Kit (Qiagen) per manufacturer’s instructions. Extracted RNA was then reverse transcribed into cDNA using the Superscript III Kit (Invitrogen) per manufacturer’s instructions. PCR for 14-3-3θ spanning the mutation site was performed with the following primers: F primer 5′-ccatcgcagagcttgataca-3′ and R primer 5′-ccaaagctgcagtgtgaaaa-3′. PCR was done using a BioRad MyCycler set to the following protocol: one cycle of denaturation at 98 °C for one minute; 30 cycles of denaturation at 98 °C for 20 s, annealing at 55 °C for 20 s, and polymerization at 68 °C for 60 s; and one cycle of extension at 68 °C for four minutes. PCR product was sent for sequencing through the UAB Heflin Sequencing core.

2.4. Preparation of conditioned media (CM)

CM was prepared as described (Wang et al., 2018). For measurement of αsyn release, isyn-derived cell lines were switched into serum-free Eagle’s MEM/F12K with doxy for 96 h prior to collection of CM. CM was treated with 5 μl of 0.1 M PMSF and then underwent serial centrifugations to remove cellular debris. CM was then concentrated using a 3 kDa Amicon Ultra-4 centrifugal filter at 4000 ×g for 120 min at 4 °C. Afterwards 5 ml 0.1 μM phenylmethylsulfonyl fluoride (PMSF) in deionized H₂0 was added to the upper device, and then the device was spun at 4000 ×g for 120 min at 4 °C. The concentrated CM was treated with protease and phosphatase inhibitors prior to BCA assay to measure protein concentration for Western blot analysis.

For toxicity experiments, isyn-derived cell lines were induced for seven days with doxy in Eagle’s MEM/F12K with 10 % FBS, then switched into serum-free Eagle’s MEM/F12K media for 48 h, and finally transferred to target cells, as previously described (Wang et al., 2018). For αsyn immunodepletion from CM in some toxicity assays, 30 μl Protein G-conjugated Dynabeads (Thermo Fisher Scientific) were incubated with mouse monoclonal anti-αsyn antibody (final concentration 1.5 μg/ml; BD Biosciences #610787) or with mouse IgG overnight prior to adding to CM. 3 ml CM was then incubated with antibody-coupled Dynabeads overnight before use for toxicity studies.

2.5. Ethidium D cell death assay

Cell death was measured as previously described (Wang et al., 2018). Cells were incubated with Ethidium D (1 μM) and Hoechst 33342 (2 μg/ml) in media for 30 min at 37 °C. Ten fields per well at 20× magnification were imaged at random for quantification. The ratio of Ethidium D-positive cells to total Hoechst 33342-positive cells were counted in each field. The rater was blinded to experimental conditions.

2.6. Autophagic flux assay

Cell lines were induced with 10 μm doxy for 96 h and then treated with 40 μM chloroquine for three hours before being lysed in 0.5 % NP-40 lysis buffer with protease and phosphatase inhibitors. Lysates were assessed for LC3b-I and -II via Western blotting.

2.7. Western blot

Primary antibodies used are listed in Table 1. Cell lysis and Western blot were performed, as previously (Wang et al., 2018). Isyn cells were lysed in lysis buffer (150 mM NaCl, 10 mM Tris-HCl, pH 7.4, 1 mM EGTA, 1 mM EDTA, 0.5 % NP-40, protease and phosphatase inhibitor cocktails; ThermoFisher Scientific) and spun at 16000 ×g for ten minutes. Supernatants from the lysates were boiled in 4× DTT sample loading buffer (0.25 M Tris-HCl, pH 6.8, 8 % SDS, 200 mM DTT, 30 % glycerol, Bromophenol Blue) prior to loading of equal protein amounts onto 12 % SDS-polyacrylamide gels. After transfer to nitrocellulose membranes followed by blocking in 5 % nonfat dry milk in TBST, membranes were incubated overnight with primary antibodies listed in Table 1 at 4 °C and then with HRP-conjugated secondary antibodies for two hours at room temperature. After several washes in TBST, enhanced chemiluminescence was used to develop the blots. Full blots are shown in Supp. Fig. 2).

Table 1.

Primary antibodies used.

Antibody	Catalog	RRID
14-3-3θ, mouse monoclonal (5J20)	Santacruz, sc-69,720	RRID: AB_2218224
14-3-3θ, mouse monoclonal (3B9)	Santa Cruz, sc59414	RRID: AB_2218234
14-3-3θ, rabbit polyclonal	Bethyl Laboratories, A303–146 A	RRID: AB_10894703
αsyn, mouse monoclonal	BD Biosciences #610787	RRID: AB_398108
αsyn, rabbit polyclonal	Cell signaling technology, #2642	RRID: AB_2192679
β-actin, mouse monoclonal	Thermo Fisher # MA1–140	RRID: AB_2536844
α-tubulin	Thermo Fisher # MA1–19162	RRID: AB_1070259
GAPDH, rabbit monoclonal	Cell signaling technology, #2118S	RRID: AB_561053
LC3B, rabbit polyclonal MAP2	Sigma #L7543	RRID: AB_796155
NeuN, rabbit monoclonal	Abcam, #ab177487	RRID: AB_2532109
phospho-S232–14-3-3θ	Invitrogen, #PA5–40283	RRID: AB_2608333
phospho-S129-αsyn, rabbit monoclonal	Abcam #ab51253	RRID: AB_869973

2.8. Immunoprecipitation

Cell lysates and CM from isyn and isyn cells expressing WT, S232A, or S232D 14-3-3θ induced for 96 h were used for immunoprecipitation, as previously described (Wang et al., 2018). 40 μl slurry of 50 % Protein G Dynabeads (Invitrogen) was incubated overnight at 4 °C with 6 μg of either rabbit anti-αsyn antibody or rabbit IgG prior to incubation with samples (750 μg). After washes in PBS with 0.02 % Tween, protein/bead complexes were boiled in DTT sample loading buffer, and then run on a 12 % SDS-PAGE gel for 14-3-3θ detection. After transfer, membranes were probed for 14-3-3θ first by using mouse anti-14-3-3θ antibody and then re-probed for αsyn using mouse anti-αsyn antibody.

2.9. S1-S2 complementation assay

Complementation assays for αsyn in the conditioned media were performed, as previously described (Danzer et al., 2011; Wang et al., 2018). Briefly, H4 cells in serum-free and phenol red-free OptiMEM media were transfected with empty vector or 14-3-3θ S232 mutants along with αsyn-hGLuc1 (syn—S1) and αsyn-hGLuc2 (syn—S2). 48 h later, the CM was centrifuged for five minutes at 3000 ×g prior to addition of 40 μM coelenterazine. A Synergy 2 plate reader (Biotek) was used to measure luciferase activity at 480 nm.

2.10. αSyn fibril preparation for primary culture experiments

Human αsyn fibrils were generated as before (Wang et al., 2018). Briefly, purified monomeric human αsyn was incubated at 37 °C with constant agitation at 700 rpm for seven days at a concentration of 5 mg/ml in 50 mM Tris (pH 7.5) with 166 mM KCl. After dilution to 1 mg/ml in PBS, αsyn was sonicated using a QSonica700 cup horn sonicator in a 16 °C water bath for 15 min at 30 % amplitude, with pulses cycling three seconds on and two seconds off. Fibril fragmentation was confirmed with dynamic light scattering (DLS) on a DynaPro NanoStar (Wyatt Technology). Sonicated fibrils were added to primary cultures at 1 μg/ml in neuronal media.

2.11. Primary neuronal culture preparation

Primary hippocampal cultures were prepared as before from male and female postnatal 0–1 (P0-P1) mice as before (Wang et al., 2018). Briefly, dissected hippocampi were incubated in papain for 25 min at 37 °C and then titurated using fire polished glass pipettes. Cells were spun done at 1000 rpm for five minutes and then plated on 18 mm poly-D-lysine-coated glass coverslips in Neurobasal-A media containing B-27 supplement and 5 % FBS. After 12–16 h, cells were switched into fresh Neurobasal-A media with B-27 supplement and Arabinose C at 6 μM.

2.12. Immunocytochemistry

Antibodies used for immunostaining are listed in Table 1. For 14-3-3θ immunostaining in brain sections, coronal brain sections were incubated in a sodium citrate antigen retrieval buffer (10 mM sodium citrate, 0.05 % Tween-20, pH 6.0), washed in TBS, and then permeabilized with 0.1 % Triton X-100. Sections were then blocked in 5 % normal goat serum (NGS) and incubated overnight in primary mouse antibodies for 14-3-3θ and NeuN. Sections were next washed and incubated at room temperature for one hour in secondary antibodies (Cy3-conjugated goat anti-mouse, Jackson Labs #115-005-003, or Alexa Fluor 488 goat anti-rabbit Invitrogen # A11008) before being washed and mounted onto slides. Slides were coverslipped using ProLong Diamond mounting media with DAPI (Thermo Fisher). Slides were imaged using an Olympus VS200 slide scanner.

For 14-3-3θ and αsyn immunostaining in culture, primary neurons were fixed in 4 % paraformaldehyde with 4 % sucrose. After permeabilization with 0.1 % Triton X-100 in PBS for ten minutes, neurons were blocked with 10 % NGS in PBS for one hour and then incubated overnight with primary antibodies in 10 % NGS and 0.1 % Triton X-100 in PBS. After washing, neurons were incubated with goat anti-rabbit or anti-mouse secondary antibodies in 10 % NGS and 0.1 % Triton X-100 in PBS. Nikon Eclipse Ti2 confocal microscope was used to image neurons.

Primary neurons were immunostaining for phospho-S129-αsyn (pS129-αsyn) after fixation in 4 % paraformaldehyde with 4 % sucrose and 1 % Triton X-100 for 15 min as previously performed (Wang et al., 2018). An Olympus BX51 epifluorescence microscope was used for imaging ten random high power (20×) fields per well for quantification. ImageJ (Schneider et al., 2012) was used to quantitate the area positive for pS129-αsyn immunoreactivity. The rater was blind to experimental conditions.

2.13. Lysosomal acid lipase (LAL) assay

Isyn cell lines were plated on 35 mm MatTek dishes and induced with doxy for 96 h. To examine LAL activity, the cells were incubated with an LAL substrate (LysoLive Lysosomal Acid Lipase Assay Kit, Abcam #ab253380) for six hours and imaged under a Zeiss Z1 Cell Observer high-speed live cell imaging system. For primary hippocampal neurons, LAL activity was measured at DIV 8.

2.14. DQ-BSA assay

Isyn cell lines were plated on 35 mm MatTek dishes and induced with doxy for 96 h. Cells were incubated with DQ-BSA substrate (DQ Red BSA, Invitrogen #D12051) for six hours prior to imaging using a Zeiss Z1 Cell Observer high-speed live cell imaging system. Similarly, primary neurons plated on 35 mm MatTek live imaging dishes were treated with DQ-BSA substrate at DIV 8 and imaged after six hours.

2.15. In vivo PFF studies

Cre control and Cre+/− S232D mice were injected with mouse recombinant αsyn PFFs (Type 1), purchased from StressMarq (#SPR-324), following the protocol described in Underwood et al. (Underwood et al., 2021) with slight modifications. Prior to the start of each day of injections, fibrils were sonicated at 4 °C in a water bath sonicator (QSonica) for 15 min of total sonication, following a protocol of three seconds on and two seconds off. Amplitude was set such that power exceeded 110 W during the sonication (A = 34–35). Fibril size was checked following sonication using DLS on a DynaPro NanoStar (Wyatt Technology), and fibril size was confirmed to be under 50 nm each day. Mice were injected with 5 μl of 2 μg/μl fibrils at a rate of 0.5 μl/min into the striatum (AP: 1.0 mm from bregma; ML: − 1.85 mm from midline; DV: − 3.0 mm below dura).

At six weeks post injection, mice were perfused with PBS followed by 4 % PFA using a forced pump system, and brains were sliced into 40 μm coronal sections using a sled microtome (Leica #SM2010R). Every sixth section through the sensorimotor regions were then stained for p-S129-αsyn), as previously described (Underwood et al., 2021). Sections were imaged using an Olympus VS200 slide scanning microscope. Using QuPath software, an ROI was drawn around the sensorimotor cortical region, somatic p-αsyn-positive inclusions were manually counted, and the number of inclusions per mm² was calculated.

2.16. Human brain lysates

The Banner Sun Health Research Institute Brain and Body Donation Program provided flash-frozen temporal cortical samples of patients with Incidental Lewy Body Disease and age-matched controls. UAB’s Institutional Review Board (IRB) determined the cortical samples were not considered human subjects because the tissue came from deceased individuals. Samples were homogenized and fractioned, as previously described (McFerrin et al., 2017). Fresh-frozen brain samples were homogenized in lysis buffer (50 mM Tris-HCl pH 7.4, 175 mM NaCl, 5 mM EDTA, protease and phosphatase inhibitors) and sonicated for 10 s on ice. Triton X-100 was added to a final concentration of 1 % v/v prior to incubation incubated on ice for 30 min. The supernatant was saved as the Triton X-100 soluble fraction after spinning at 15,000 ×g for one hour at 4 °C. The pellet was resuspended in lysis buffer with 2 % SDS, sonicated for 10 s, and centrifuged for five minutes at 15,000 ×g in order to prepare the Triton X-100 insoluble fraction.

Samples were run on duplicate 26-well precast 12 % SDS-PAGE gels simultaneously. One gel was assessed for total protein using GelCode Blue Safe protein stain (Thermo Fisher Scientific #1860957). The other gel was transferred onto a PVDF membrane and probed for phospho-S232 and total 14-3-3θ, as previously described (McFerrin et al., 2017). Total 14-3-3θ and phospho-S232 were imaged using fluorescent secondary antibodies on a LICOR Odyssey Imaging system.

2.17. Statistical analysis

Graph Pad Prism 10 was used for statistical analysis. Experiments involving two groups were analyzed by two-tailed, unpaired student’s t-test or by two-tailed Mann Whitney test. Experiments involving more than two groups were analyzed by 1-way or 2-way ANOVA, followed by post-hoc Dunnett’s, Sidak’s, or Tukey’s multiple comparison test. Details of statistical analyses for each experiment are found in Supp. Table 1.

3. Results

3.1. Phosphorylation blocks 14-3-3θ’s protective effects against released αsyn

Our lab has previously developed a paracrine doxy-inducible αsyn cell model, called isyn, which allows us to examine the cellular mechanisms regulating αsyn transmission from one cell to another (Wang et al., 2018). Upon doxy treatment, isyn cells overexpress human αsyn which is then released into the conditioned media (CM) primarily through non-exosomal means (Wang et al., 2018). This αsyn-enriched CM induces toxicity when transferred to separately cultured cells, and αsyn depletion demonstrates that this toxicity is mediated by αsyn (Wang et al., 2018). We previously observed that overexpression of 14-3-3θ reduced αsyn toxicity in this paracrine model, while 14-3-3 inhibition exacerbated this toxicity (Wang et al., 2018).

Here we examined the impact of 14-3-3θ phosphorylation at S232 in this model. Isyn cells were transduced with a V5-tagged WT, S232A, or S232D 14-3-3θ doxy-inducible lentiviruses and then selected for polyclonal stable transfection. Separately cultured differentiated SH-SY5Y cells showed ~17 % cell death at 48 h after treatment with CM from induced isyn cells, but toxicity was back to baseline when treated with CM from induced WT 14-3-3θ-overexpressing isyn cells (Fig. 1a, b), as previously observed (Wang et al., 2018). Similarly, the non-phosphorylatable S232A 14-3-3θ mutant reduced the toxicity of αsyn-enriched CM back to baseline (Fig. 1a, b). In contrast, the phosphomimetic S232D 14-3-3θ mutant did not protect against αsyn-enriched CM toxicity (Fig. 1a, b). While intracellular 14-3-3θ levels were increased in isyn cells expressing V5-tagged WT or mutant 14-3-3θ compared to isyn cells, we did not observe differences in endogenous or exogenous 14-3-3θ expression among isyn cells expressing WT, S232A, or S232D 14-3-3θ (Fig. 2b). Therefore, the differential effects of 14-3-3θ mutants on αsyn toxicity were not related to differences in 14-3-3θ expression levels.

Fig. 1.

14-3-3θ phosphorylation at S232 prevents its ability to protect against paracrine αsyn toxicity. a. Representative images of differentiated SH-SY5Y cells treated with conditioned media (CM) from doxy-induced isyn cells, isyn cells expressing WT 14-3-3θ, isyn cells expressing the non-phosphorylatable S232A 14-3-3θ mutant, and isyn cells expressing the phosphomimetic S232D 14-3-3θ mutant. Ethidium D (EthD) stains nuclei of dying cells. Hoechst 33342 stains nuclei of all cells. Scale bar 100 μm. b. Quantification of cell death in differentiated SH-SY5Y cells treated with αsyn-enriched CM from uninduced and induced isyn cells with or without WT or mutant 14-3-3θ. n = 3 independent rounds with two replicates per round. ^***p < 0.001, ^****p < 0.0001 (Tukey’s multiple comparison test). c. Quantification of cell death in differentiated SH-SY5Y cells treated with CM from induced isyn cells with and without WT or mutant 14-3-3θ after immunodepletion for αsyn. n = 3 independent rounds with two replicates per round. ^**p < 0.01, ^****p < 0.0001 (Sidak’s multiple comparison test). n.s. non-significant. d. Western blot analysis of CM after pulldown of αsyn with a monoclonal antibody against αsyn demonstrated reduction of αsyn levels upon immunodepletion. Error bars represent standard error.

Fig. 2.

14-3-3θ phosphorylation at S232 does not affect the release of αsyn into the conditioned media. a. Representative Western blots and quantification of αsyn, endogenous 14-3-3θ, and exogenous V5-tagged 14-3-3θ in the CM from induced isyn cells, isyn cells expressing WT 14-3-3θ, isyn cells expressing the S232A mutant, and isyn cells expressing the S232D mutant. Exogenous, V5-tagged WT or mutant 14-3-3θ (marked by arrowhead) runs a little higher than the endogenous 14-3-3θ band due to the V5 epitope tag. n = 6 independent rounds. * p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001 (Tukey’s multiple comparison test). b. Representative Western blots and quantification of αsyn, endogenous 14-3-3θ, and exogenous V5-tagged 14-3-3θ in the lysates from induced isyn cells, isyn cells expressing WT 14-3-3θ, isyn cells expressing the S232A mutant, and isyn cells expressing the S232D mutant. Exogenous, V5-tagged WT or mutant 14-3-3θ (marked by arrowhead) runs a little higher than the endogenous 14-3-3θ band due to the V5 epitope tag. n = 4 independent rounds. * p < 0.05, ^** p < 0.01 (Tukey’s multiple comparison test). Error bars represent standard error.

To test whether paracrine toxicity induced by CM from isyn cells expressing the S232D 14-3-3θ mutation was mediated by αsyn, we performed αsyn immunodepletion from the CM prior to transfer to separately cultured cells. As previously observed (Wang et al., 2018), αsyn immunodepletion of the CM from doxy-induced isyn cells reduced cell death near control levels (Fig. 1c). Similarly, CM from doxy-induced isyn cells expressing S232D mutant also showed dramatic reduction in toxicity upon αsyn immunodepletion (Fig. 1c). αSyn immunodepletion was confirmed by Western blot analysis (Fig. 1d).

3.2. 14-3-3θ phosphorylation does not affect αsyn levels

We next examined the impact of 14-3-3θ phosphorylation on αsyn intracellular levels and on αsyn release from isyn cells. Total intracellular αsyn levels were not altered in isyn cells with expression of either WT or S232 mutant forms of 14-3-3θ (Fig. 2b). As we have previously seen (Wang et al., 2018), overexpression of WT 14-3-3θ did increase the total amount of αsyn released into the CM by nearly three-fold (Fig. 2a; Supp. Fig. 3). Expression of the S232A mutant or the S232D mutant similarly increased the total amount of αsyn released (Fig. 2a; Supp. Fig. 3).

We also observed release of endogenous 14-3-3θ into the CM with isyn cells expressing WT, S232A, or S232D 14-3-3θ: higher levels of endogenous 14-3-3θ were released into the CM in isyn cells expressing WT, S232A, or S232D 14-3-3θ compared to that from isyn cells alone (Fig. 2a). Exogenous V5-tagged 14-3-3θ was also released into the CM, but at much lower amounts compared to the endogenous 14-3-3θ as we have previously observed (Wang et al., 2018). Interestingly, less mutant S232A or S232D 14-3-3θ was released into the CM compared to V5-tagged WT 14-3-3θ. Differences in 14-3-3θ intracellular expression levels did not account for the differential effect of 14-3-3θ mutants on 14-3-3θ release: we found that the levels of exogenous and endogenous WT and mutant 14-3-3θ intracellular expression were comparable among isyn cells expressing WT, S232A, or S232D 14-3-3θ (Fig. 2b).

3.3. Released αsyn has differential oligomerization upon 14-3-3θ phosphorylation

Oligomerization of αsyn is critical to αsyn toxicity. We have previously shown that released αsyn in the paracrine model demonstrates oligomeric features, and that 14-3-3θ overexpression reduces αsyn oligomerization upon binding (Wang et al., 2018). To test the impact of S232 phosphorylation on the ability of 14-3-3θ to reduce αsyn oligomerization, we used a bioluminescent protein-fragment complementation assay. In this assay, a luminescent signal is generated when αsyn fused to the N-terminal non-bioluminescent luciferase fragment (S1-syn) oligomerizes with αsyn fused to the C-terminal luciferase fragment (S2-syn) in H4 glioma cells (Danzer et al., 2011). We previously showed that WT 14-3-3θ reduces the luciferase signal in the media of H4 cells transfected with S1-syn and S2-syn (Wang et al., 2018). Similarly, S232A 14-3-3θ expression reduced the luciferase signal in the media from H4 cells transfected with S1-syn and S2-syn, but S232D 14-3-3θ expression did not significantly reduce the luciferase signal in the CM (Fig. 3).

Fig. 3.

The S232A mutant reduces oligomerization of αsyn in the CM compared to the S232D mutant. Quantification of luciferase activity in media from H4 neuroglioma cells transfected with S1-syn, S2-syn, and empty vector (EV), S232A, or S232D 14-3-3θ. n = 3 independent rounds with 2–3 replicates/rounds. ^** p < 0.01, ^*** p < 0.001 (Tukey’s multiple comparison test). Error bars represent standard error.

3.4. S232 phosphorylation does not impact autophagic-lysosomal pathways

Autophagic-lysosomal pathways are critical to the clearance and toxicity of αsyn. 14-3-3θ has been shown to regulate autophagic-lysosomal pathways in non-neuronal cells (Wang et al., 2010). Here we tested whether autophagic-lysosomal function is affected by 14-3-3θ phosphorylation. Control isyn cells and isyn cells expressing either WT, S232A, and S232D 14-3-3θ were induced with doxy for 96 h prior to chloroquine (CQ) treatment (40 μM) for three hours to measure autophagic flux. CQ treatment promotes buildup of LC3 II through inhibition of autophagosome/lysosome fusion. As expected, LC3 II levels were increased after CQ treatment in induced isyn cells (Fig. 4a). Similar elevations in LC3 II levels with CQ treatment were seen in induced isyn cells expressing WT, S232A, or S232D 14-3-3θ (Fig. 4a).

Fig. 4.

Lysosomal function is not affected by WT or mutant 14-3-3θ expression in isyn cells. a. Representative blots and quantification of LC3-I and LC3-II levels after CQ treatment in doxy-induced isyn cells without and with WT or mutant 14-3-3θ expression. n = three independent rounds. * p < 0.05 (Tukey’s multiple comparison test). n.s. non-significant. b. Representative images and quantification of DQ-BSA activity in doxy-induced isyn cells without and with WT or mutant 14-3-3θ expression. n = 10 individual cells per condition per round with three independent rounds (total 30 cells imaged per condition). n.s. (One-way ANOVA). Scale bar 20 μm. c. Representative images and quantification of lysosomal lipase A (LAL) activity in doxy-induced isyn cells without and with WT or mutant 14-3-3θ expression. n = 10 individual cells per condition per round with three independent rounds (total 30 cells imaged per condition). n.s. (One-way ANOVA). Scale bar 20 μm. Error bars represent standard error.

We also measured lysosomal proteolytic activity using the DQ-BSA assay in isyn cells with or without 14-3-3θ overexpression upon doxy induction. DQ-BSA cleavage was unchanged in doxy-induced isyn/WT 14-3-3θ cells compared to doxy-induced isyn cells (Fig. 4b). There were also no differences in DQ-BSA cleavage among induced isyn cells expressing either WT, S232A, or S232D 14-3-3θ (Fig. 4b). We also measured lysosomal acid lipase (LAL) activity in induced isyn cells with and without WT or mutant 14-3-3θ expression to validate these findings. LAL activity levels were comparable in induced isyn cells and in induced isyn cells expressing either WT, S232A, and S232D 14-3-3θ (Fig. 4c). Based on these data, autophagic-lysosomal function was not impacted by WT or mutant 14-3-3θ expression in isyn cells.

3.5. S232 phosphorylation disrupts 14-3-3θ’s interaction with αsyn

We next evaluated whether S232 phosphorylation affects 14-3-3θ’s ability to interact with αsyn. V5-tagged WT 14-3-3θ co-immunoprecipitated with αsyn in cellular lysates from induced isyn cells with 14-3-3θ overexpression (Fig. 5a). Similarly, V5-tagged S232A 14-3-3θ co-immunoprecipitated with αsyn from cell lysates from induced isyn cells with S232A mutant expression, but V5-tagged S232D failed to co-immunoprecipitate with αsyn from cell lysates from induced isyn cells with S232D mutant expression (Fig. 5a). In the CM, we confirmed that endogenous 14-3-3θ bound to released αsyn in the CM from isyn cells with 14-3-3θ overexpression, as previously seen (Wang et al., 2018) (Fig. 5b). Expression of S232A in isyn cells did not impact the amount of endogenous 14-3-3θ bound to released αsyn, but S232D expression in isyn cells dramatically reduced the amount of endogenous 14-3-3θ that co-immunoprecipitated with αsyn from the CM (Fig. 5b). Minimal interaction of the V5-tagged 14-3-3θ proteins was detected in the CM after αsyn immunoprecipitation (Fig. 5b). This effect of S232D mutant on endogenous 14-3-3θ co-immunoprecipitation from the CM suggests that the S232D phosphomimetic mutation acts as a dominant negative to interfere with endogenous 14-3-3θ binding to αsyn in the CM.

Fig. 5.

14-3-3θ phosphorylation reduces 14-3-3θ’s binding to αsyn. a. Representative blot and quantification of co-immunoprecipitation of exogenous (V5-tagged) 14-3-3θ with αsyn from cell lysates of induced isyn cells without or with WT or mutant 14-3-3θ expression. n = 4 independent rounds. ^**p < 0.01, ^****p < 0.0001 (Tukey’s multiple comparison test). b. Representative blot and quantification of co-immunoprecipitation of endogenous 14-3-3θ with αsyn from CM of induced isyn cells without or with WT or mutant 14-3-3θ expression. n = 4 independent rounds. ^*****p < 0.0001 (Tukey’s multiple comparison test). Error bars represent standard error.

3.6. Creation of novel S232 mutant mouse lines

Our studies in the paracrine model reveal that 14-3-3θ phosphorylation at S232 inhibits the ability of 14-3-3θ to reduce αsyn binding, oligomerization, and toxicity – suggesting that 14-3-3θ phosphorylation disrupts 14-3-3θ’s chaperone function. To further assess the impact of phosphorylation on 14-3-3θ function, we developed novel conditional knock-in (KI) mice that express either the non-phosphorylatable S232A mutant or the phosphomimetic S232D mutant of 14-3-3θ upon Cre-mediated recombination (Supp. Fig. 4a-c). We crossed the conditional S232A mutant mice (Cre −/− S232A +/+) with heterozygous Emx1-Cre mice (Cre +/− S232A −/−) (Gorski et al., 2002) to induce S232A expression into cortical and hippocampal regions of the brain. These heterozygous S232A mice (Cre +/− S232A +/−) were then crossed with Cre −/− S232A +/− mice to generate the experimental mice needed: Cre control mice (Cre +/− S232A −/−) and homozygous S232A KI mice (Cre +/− S232A +/+) (Supp. Fig. 4d). Cre +/− S232A +/+ mice are viable, and we confirmed that mRNA for the mutant S232A is expressed in these mice (Supp. Fig. 4d). 14-3-3θ immunoreactivity was similar between Cre control mice and Cre +/− S232A +/+ mice (Supp. Fig. 4f).

Similarly, we crossed the conditional S232D mutant mice (Cre −/− S232D +/+) with Emx1-Cre mice (Cre +/− S232D −/−) (Gorski et al., 2002) to induce S232D expression in the cortex and hippocampus. Cre +/− S232D +/− mice were then crossed with Cre −/− S232D +/− mice to generate the following: WT mice (Cre −/− S232D −/−); Cre control mice (Cre +/− S232D −/−); heterozygous S232D KI mice (Cre +/− S232D +/−); and homozygous S232D KI mice (Cre +/− S232D +/+) (Supp. Fig. 4e). Cre +/− S232D +/+ mice are viable, and mRNA for the mutant S232D is expressed in these mice (Supp. Fig. 4e). 14-3-3θ immunoreactivity was similar between Cre control and Cre +/− S232D +/+ mice (Supp. Fig. 4f).

3.7. 14-3-3θ phosphorylation does not affect protein levels of 14-3-3θ or αsyn

We examined whether 14-3-3θ phosphorylation affected the protein levels or subcellular localization of 14-3-3θ or αsyn. We first measured 14-3-3θ and αsyn levels in Cre control and S232D mouse brain lysates. We observed no differences in 14-3-3θ levels in Triton X-100 soluble and insoluble fractions from hippocampal lysates from three month old Cre control and homozygous S232D KI mice (Fig. 6a). αSyn levels were similar in the soluble or insoluble fractions in hippocampal lysates from Cre control and homozygous S232D KI mice (Fig. 6a). In primary hippocampal cultures from WT and homozygous S232D KI mice, immunoreactivity for both 14-3-3θ and αsyn showed similar cellular distribution (Fig. 6b).

Fig. 6.

αSyn levels are not altered in mutant S232 KI mouse brains. a. Representative Western blot and quantification of αsyn and 14-3-3θ levels in Triton X-100 soluble and insoluble fractions from the hippocampus of 3–4 month old Cre control and Cre+/− S232D+/+ mice. n.s. (Student’s t-test). b. Immunocytochemistry shows similar distribution of 14-3-3θ and αsyn in primary hippocampus cultures from wildtype (Cre−/− S232D−/−) and Cre+/− S232D+/+ mice. Scale bar 50 μm. c. Representative Western blot and quantification of αsyn and 14-3-3θ levels in Triton X-100 soluble and insoluble fractions from the hippocampus of 3–4 month old Cre control and Cre+/− S232A+/+ mice. n.s. (Student’s t-test). d. Immunocytochemistry shows similar distribution of 14-3-3θ and αsyn in primary hippocampus cultures from Cre control and Cre+/− S232A+/+ mice. Scale bar 50 μm. Error bars represent standard error.

We also examined 14-3-3θ and αsyn expression in the S232A KI mice. Three month old Cre control and homozygous S232A KI mice showed similar expression levels for 14-3-3θ and αsyn in Triton X-100 soluble and insoluble lysates from the hippocampus (Fig. 6c). Immunoreactivity showed similar cellular distribution for both 14-3-3θ and αsyn in WT and S232A hippocampal cultures (Fig. 6d).

3.8. S232 phosphorylation accelerates αsyn aggregation in the PFF model

We next examined the effect of 14-3-3θ phosphorylation in the preformed fibril (PFF) culture model (Volpicelli-Daley et al., 2011). As previously described, PFFs induce aggregation of endogenous mouse αsyn in primary neuronal cultures (Volpicelli-Daley et al., 2011). We have previously observed that this aggregation induced by αsyn PFFs is dramatically reduced in hippocampal cultures from transgenic 14-3-3θ mice compared to WT mice (Wang et al., 2018). To test the impact of 14-3-3θ phosphorylation, we treated primary hippocampal neurons from Cre control mice, mice heterozygous for S232D (Cre +/− S232D +/−), and mice homozygous for S232D (Cre +/− S232D +/+) with human monomeric or PFF αsyn (1 μg/ml) at DIV 5 and then measured insoluble, phosphorylated S129-αsyn (pS129-αsyn) immunostaining as a measure of pathologic αsyn inclusions at 7, 10, and 14 days after treatment.

Treatment of hippocampal neurons from Cre control mice showed an increase in phosphorylated pS129-αsyn immunostaining in response to human αsyn PFF treatment (Fig. 7a, b). pS129-αsyn immunoreactivity was detectable at seven days after PFF treatment and increased over time in Cre control neurons (Fig. 7a, b). In hippocampal cultures from mice heterozygous for S232D (Cre +/− S232D +/−), the amount of pS129-αsyn staining was significantly increased at both 10 and 14 days after PFF treatment compared to Cre control cultures (Fig. 7a, b). pS129-αsyn levels were greater in hippocampal cultures from mice homozygous for S232D (Cre +/− S232D +/+) compared to either Cre control or heterozygous S232D cultures treated with αsyn PFFs at 10 days after treatment, suggesting a dose-dependent increase (Fig. 7a, b). Neurons from mice homozygous for S232D (Cre +/− S232D +/+) also showed increased pS129-αsyn immunostaining compared to Cre control at 14 days after PFFs (Fig. 7a, b). Monomeric αsyn treatment of neurons from Cre control, Cre +/− S232D +/−, or Cre +/− S232D +/+ mice did not induce αsyn aggregation, as assessed by pS129-αsyn immunoreactivity (Fig. 7a).

Fig. 7.

The S232D mutation increases αsyn inclusion formation, while the S232A mutation protects against αsyn inclusion formation upon PFF treatment in neurons. a. Representative images of Cre control, Cre+/− S232D+/−, and Cre+/− S232D+/+ hippocampal neurons treated with monomeric or fibrillary αsyn at DIV 5 and then fixed and stained at 7, 10, or 14 days after treatment for insoluble, pS129 αsyn. Scale bar 100 μm. b. Quantification of area positive for pS129-αsyn immunostaining in hippocampal cultures from Cre control, Cre+/− S232D+/−, and Cre+/− S232D+/+ treated with αsyn PFFs. n = six independent rounds with 1–2 replicates per round. *p < 0.05, ^****p < 0.0001 (Tukey’s multiple comparison test). c. Quantification of DQ-BSA activity in Cre control and Cre+/− S232D+/+ primary hippocampal neurons at DIV 8. n.s. (Student’s t-test). d. Quantification of LAL activity in Cre control and Cre+/− S232D+/+ primary hippocampal neurons at DIV 8. n.s. (Student’s t-test). e. Representative images of Cre control and Cre+/− S232A+/+ hippocampal neurons treated with fibrillary αsyn at DIV5 and then fixed and stained for insoluble, pS129 αsyn at 7, 10, or 14 days after treatment. Scale bar 100 μm. f. Quantification of area positive for pS129-αsyn immunostaining in hippocampal cultures from Cre control and Cre+/− S232A+/+ treated with αsyn PFFs. n = three independent rounds with two replicates per round. *p < 0.05, ^****p < 0.0001 (Tukey’s multiple comparison test). Error bars represent standard error.

We hypothesized that increased αsyn aggregation could be secondary to disruption of lysosomal function in Cre +/− S232D +/+ neurons. To test if lysosomal function is affected by 14-3-3θ phosphorylation in primary neurons, we measured cleavage of DQ-BSA in Cre control and Cre +/− S232D +/+ neurons at DIV 8. DQ-BSA cleavage levels were comparable in Cre control and Cre +/− S232D +/+ neurons (Fig. 7c). Similarly, LAL activity was comparable in Cre control and Cre +/− S232D +/+ neurons (Fig. 7d).

Similarly, we tested the effect of PFF treatment in S232A-expressing neurons. We found that Cre +/− S232A +/+ neurons showed ~20 % reduction of pS129-αsyn staining upon PFF treatment compared to Cre control neurons at 10 and 14 days after treatment (Fig. 7e, f). Monomeric αsyn treatment of Cre control and Cre +/− S232A+/+ neurons did not show any pS129-αsyn immunoreactivity (Supp. Fig. 5).

We next tested the impact of 14-3-3θ phosphorylation on αsyn aggregation in vivo. Cre control and Cre +/− S232D +/+ mice were injected with mouse αsyn PFFs into the dorsolateral striatum at 8–12 weeks of age. At six weeks post injection, we measured αsyn aggregation by pS129-αsyn immunohistochemistry in the sensorimotor cortex (Supp. Fig. 6). Control mice showed extensive inclusion formation in the sensorimotor cortex (Fig. 8). Cre +/− S232D +/+ mice showed a 27 % increase in inclusion counts in the cortex compared to Cre control mice (Fig. 8).

Fig. 8.

αSyn PFFs induce increased αsyn inclusions in the cortex of S232D mice. a. Representative images of pS129 αsyn immunoreactivity in cortical sections from Cre control and Cre+/− S232D+/+ mice six weeks after αsyn PFF injections into the dorsolateral striatum. Scale bar 100 μm. b. Quantification of pS129-αsyn positive inclusions per mm² in the sensorimotor cortex of Cre control and Cre+/− S232D+/+ mice. *p < 0.05 (Student’s t-test). Error bars represent standard error.

3.9. 14-3-3θ phosphorylation at S232 in human incidental lewy body disease

We previously have observed that 14-3-3θ phosphorylation at S232 is increased in Triton X-100 insoluble fractions from human PD and DLB lysates from temporal cortex compared to age-matched controls (McFerrin et al., 2017). To determine if these changes occur in the pre-symptomatic stage, we examined 14-3-3θ phosphorylation in brain lysates from deceased subjects pathologically diagnosed with incidental Lewy Body Disease (iLBD). Pathological diagnosis of iLBD is defined by the presence of Lewy neurite or Lewy bodies in autopsied subjects without clinical features of cognitive decline or motor signs of Parkinsonism prior to death (DelleDonne et al., 2008; Dickson et al., 2008; Dijkstra et al., 2014; Iacono et al., 2015). 14-3-3θ phosphorylation at S232 in Triton X-100 soluble and insoluble fractions did not differ between control and iLBD subjects (Fig. 9). Interestingly, iLBD cortical lysates showed a small reduction in insoluble 14-3-3θ levels in the Triton X-100 insoluble fractions compared to control lysates. These data suggest that 14-3-3θ phosphorylation changes occur later in the disease process.

Fig. 9.

14-3-3θ phosphorylation is not observed in incidental Lewy Body Disease. a. Representative Western blots of phospho-S232 14-3-3θ (pS232) and total 14-3-3θ in Triton X-100 soluble and insoluble fractions from temporal cortex of control and iLBD human brains. b. Quantification of pS232 and total 14-3-3θ levels in control and iLBD cortex. n = 42 individual brain samples for control; n = 44 individual brain samples for iLBD. Error bars represent standard error. *p < 0.05 (Mann Whitney or student t-test).

4. Discussion

We examined the impact of 14-3-3θ phosphorylation on αsyn aggregation in several different αsyn models. Using a paracrine inducible αsyn model, we found that the non-phosphorylatable S232A 14-3-3θ mutant reduced the toxicity of αsyn-enriched CM similarly to WT 14-3-3θ, yet the phosphomimetic S232D mutant did not protect against paracrine αsyn toxicity. Both WT and S232A 14-3-3θ were able to bind to intracellular αsyn, while S232D showed reduced co-immunoprecipitation with intracellular αsyn. In the CM fraction, S232D reduced the interaction of αsyn with endogenous 14-3-3θ, pointing to a potential dominant negative effect of the S232D mutant. Using new S232 mutant KI mouse models, we found that S232D expression promoted αsyn aggregation in vitro and in vivo in response to PFF treatment. In contrast, S232A expression reduced αsyn aggregation. These findings suggest that S232 phosphorylation disrupts the chaperone function of 14-3-3θ that normally prevents αsyn aggregation.

The interaction of 14-3-3θ with αsyn is likely critical for the protective effects of 14-3-3θ against αsyn aggregation and toxicity (Wang et al., 2018). Our data here suggest that 14-3-3θ phosphorylation disrupts this interaction to promote αsyn aggregation. We have previously shown that with 14-3-3θ overexpression, more 14-3-3θ is bound to αsyn which reduces its oligomeric form and its toxicity, while inhibition of 14-3-3 s with the competitive peptide inhibitor difopein reduces that interaction between endogenous 14-3-3θ and αsyn and increases αsyn oligomerization and toxicity (Wang et al., 2018). Here we see a similar finding with the S232D mutant. The S232D mutant showed reduced interaction with αsyn compared to either WT or S232A 14-3-3θ within cellular lysates. Additionally, the S232D mutant reduced the interaction of untagged, endogenous 14-3-3θ with αsyn released into the CM – suggesting that the S232D may dimerize with endogenous 14-3-3θ to act as a dominant negative to reduce this interaction. 14-3-3 proteins normally mediate their interactions primarily as dimers (Mackintosh, 2004; Pair and Yacoubian, 2021). This reduction in αsyn binding in isyn cells expressing the S232D mutant was associated with a lack of protection against αsyn toxicity, while both WT and S232A 14-3-3θ led to protection against paracrine αsyn toxicity. Using an αsyn complementation assay, we found that S232D 14-3-3θ failed to reduce oligomeric αsyn in the CM compared to S232A mutant. These findings, along with the effects of S232D 14-3-3θ in the PFF model, point to a disruption of 14-3-3θ chaperone function upon S232 phosphorylation. Further studies using biophysical approaches need to examine how S232D phosphorylation affects 14-3-3θ structure and 14-3-3θ’s interaction with αsyn. Molecular modeling by our group has shown that phosphorylation at S232 causes a 90 degree kink in the C-terminal helix that may impact the amphipathic binding pocket to destabilize certain binding interactions (Pattanayak et al., 2024).

One consistent finding we have observed in the study here and previously (Wang et al., 2018) is that the endogenous 14-3-3θ protein is released into the CM at higher levels than the epitope-tagged WT or mutant 14-3-3θ protein. One cause for this lower level of exogenous 14-3-3θ release is that this exogenous 14-3-3θ is expressed at slightly lower levels than endogenous 14-3-3θ, but not to the degree to explain the larger difference in release into the CM. Another possibility is that the V5/his epitope tag may either reduce the secretion of exogenous 14-3-3θ, or alternatively, promote degradation of the exogenous 14-3-3θ in the conditioned media. Additionally, we observed that both the S232A or S232D mutants showed reduced secretion, suggesting that modification at the C-terminal end of 14-3-3θ impaired secretion.

14-3-3 proteins have multiple cellular roles besides chaperone-like activities that could contribute to the findings observed here. The 14-3-3θ isoform promotes autophagic-lysosomal clearance in non-neuronal cells (Wang et al., 2010). We examined whether autophagic flux or lysosomal function is affected by WT or mutant 14-3-3θ, but we did not observe any effects of 14-3-3θ on autolysosomal activity in either isyn cells or in primary neuronal cultures. Disruption of the interaction with other binding partners could also contribute to our findings here. Given the broad array of binding partners described for 14-3-3 proteins (Segal et al., 2023), many potential other binding interactions could be affected by 14-3-3 phosphorylation. As one example, we recently observed that 14-3-3θ phosphorylation promotes LRRK2 kinase activity through alteration in LRRK2 protein conformation (Pattanayak et al., 2024); this finding points to the possibility that activation of LRRK2 kinase activity could contribute to the findings observed here.

Examination of αsyn release and toxicity in our paracrine model revealed potentially conflicting findings. While WT and S232A 14-3-3θ protected against αsyn paracrine toxicity, they both promoted the amount of total αsyn released into the CM compared to isyn cells. This is similar to what we have previously observed in the model: while 14-3-3θ overexpression increased total αsyn levels in the CM, it reduced toxicity due to a reduction in the fraction of αsyn that was oligomeric and toxic (Wang et al., 2018). In contrast, 14-3-3 inhibition reduced total αsyn levels in the CM but increased toxicity due to an increase in the fraction of αsyn that was oligomeric and toxic (Wang et al., 2018). It is the type of αsyn species that is released and not the total amount of αsyn released that is associated with toxicity. Our complementation assay is consistent with our findings: S232A expression reduces the amount of oligomeric αsyn released into the CM compared to S232D. Additionally, we also observed an increase in the amount of endogenous 14-3-3θ released into the CM with either WT or S232A 14-3-3θ overexpression, whereas less endogenous 14-3-3θ was released with S232D expression. These data suggest that the amount of 14-3-3-θ available to interact with αsyn in the CM is also critical in determining αsyn toxicity.

We have previously observed that 14-3-3θ phosphorylation at S232 is increased in insoluble temporal cortical lysates from human PD and DLB subjects compared to age-matched controls (McFerrin et al., 2017). Our findings here suggest that this increase in 14-3-3θ phosphorylation observed in human disease contributes to αsyn aggregation and toxicity found in these disorders. We hypothesize that 14-3-3θ phosphorylation promotes the insolubilization of 14-3-3θ so that 14-3-3θ is no longer available to act as a chaperone to prevent αsyn misfolding or to participate in its other biological functions. Future studies will investigate the impact of 14-3-3θ phosphorylation on 14-3-3θ solubility.

What initiates 14-3-3θ phosphorylation in disease is not clear at this time. Our previous work suggests that exposure to pesticides like rotenone may induce 14-3-3θ phosphorylation at S232 (Slone et al., 2015). Additionally, we have previously observed that increased αsyn levels induce S232 phosphorylation (Slone et al., 2015). Of note, we did not see an increase in S232 phosphorylation in the brains of human subjects with iLBD. This suggests that S232 phosphorylation is not an early event in the disease process, yet as αsyn accumulates over the disease course, this accumulation could serve as a trigger for S232 phosphorylation. Consistent with this interpretation, we previously found that induction of αsyn levels promoted 14-3-3θ phosphorylation at S232 (Slone et al., 2015).

If 14-3-3θ phosphorylation promotes toxicity in synucleinopathies, then this phosphorylation site could serve as a therapeutic target. Future examination of the S232A mutation in in vivo PD models is indicated to assess whether blocking 14-3-3θ phosphorylation could be protective in vivo. Which kinases may be involved in phosphorylation at S232 are not known at this time. We previously showed that casein kinase 1 and casein kinase 2 inhibitors could reduce S232 phosphorylation in response to rotenone treatment in vitro (Slone et al., 2015), yet these kinases are ubiquitous and would be difficult to target chronically. BCR kinase has been shown to phosphorylate at the S232 site in vitro (Clokie et al., 2005). Other potential kinases that show consensus sequences for the S232 site include MAP kinase-activated protein kinase 2, MAP kinase-activated protein kinase 3, and polo-like kinase 1, among others (phosphonet.ca; (Safaei et al., 2011)). Future studies would examine which kinases are involved in 14-3-3θ phosphorylation at S232. Alternative methods to promote the interaction of 14-3-3θ with αsyn include 14-3-3 protein-protein interaction (PPI) stabilizers, which are currently in development as therapies for spinal cord injury (Andlovic et al., 2023; Kaplan et al., 2020; Kaplan et al., 2017; Kenanova et al., 2023; Kuusk et al., 2020; Pair and Yacoubian, 2021; Soini et al., 2022). Fusicoccin-A (FC-A), a non-specific 14-3-3 PPI stabilizer, has been shown to be protective in the AAV-αsyn mouse model (Vinueza-Gavilanes et al., 2023). It is feasible that 14-3-3 PPI stabilizers could serve to augment the 14-3-3θ/αsyn PPI destabilized by 14-3-3θ phosphorylation.

Limitations of our studies includes the use of phosphorylation mutants to test out the impact of phosphorylation of 14-3-3θ in αsyn-based models. The S232D mutation does not fully mimic phosphorylated serine residue, yet at this time this is the best approach to understanding the impact of phosphorylation in cellular and animal models. Techniques such as genetic code expansion (Qin and Liu, 2022) need to be optimized for use in mammalian systems. Additionally, in our paracrine model, we used an overexpression paradigm to test the influence of 14-3-3θ phosphorylation on paracrine αsyn toxicity. While this confirmed that 14-3-3 phosphorylation impacts the protective effects of 14-3-3θ in this model, it may be less relevant to human disease due to overexpression. However, our studies in the PFF model used a KI system in which mutant 14-3-3 levels were comparable to endogenous 14-3-3θ levels. Our finding that the S232D mutant at endogenous expression levels promoted αsyn aggregation points to the relevance of 14-3-3θ phosphorylation to human synucleinopathies.

In conclusion, we observed that 14-3-3θ phosphorylation disrupts the protective effects of 14-3-3θ against αsyn aggregation. Targeting 14-3-3θ phosphorylation with kinase inhibitors could serve as a therapeutic target.

Data availability

The data sets acquired during the current study are available from the corresponding author upon request.