Gaucher Disease Glucocerebrosidase

and a-Synuclein Form a Bidirectional

Pathogenic Loop in Synucleinopathies
Joseph R. Mazzulli,1 You-Hai Xu,2,3 Ying Sun,2,3 Adam L. Knight,4 Pamela J. McLean,1 Guy A. Caldwell,4,5
Ellen Sidransky,6 Gregory A. Grabowski,2,3 and Dimitri Krainc1,*
1Department of Neurology, Massachusetts General Hospital, Harvard Medical School, MassGeneral Institute for Neurodegenerative Disease,

Charlestown, MA 02129, USA

2The Division of Human Genetics, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229, USA
3Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH 45229, USA
4Department of Biological Sciences, University of Alabama, Tuscaloosa, AL 35487, USA
5Departments of Neurology and Neurobiology, Center for Neurodegeneration and Experimental Therapeutics, University of Alabama at

Birmingham, Birmingham, AL 35294, USA

6Section on Molecular Neurogenetics, Medical Genetics Branch, National Human Genome Research Institute, National Institutes of Health,

Bethesda, MD 20892, USA

*Correspondence: [email protected]
DOI 10.1016/j.cell.2011.06.001

SUMMARY nopathies are all characterized by the presence of aggregated,

insoluble a-syn within Lewy bodies and Lewy neurites of the
Parkinson’s disease (PD), an adult neurodegenera- central nervous system in the form of typical amyloid fibrils (Tro-
tive disorder, has been clinically linked to the lyso- janowski and Lee, 2002). The identification of PD-causing muta-
somal storage disorder Gaucher disease (GD), but tions in a-syn, which accelerate aggregation in vitro (Conway
the mechanistic connection is not known. Here, we et al., 1998) and in transgenic mice (Chandra et al., 2005; Gias-
show that functional loss of GD-linked glucocerebro- son et al., 2002), indicates that the formation of fibrils is an impor-
tant pathogenic event. Recent evidence using in vitro systems
sidase (GCase) in primary cultures or human iPS
has also indicated that soluble oligomeric a-syn assemblies,
neurons compromises lysosomal protein degrada- intermediates in fibril formation, can be cytotoxic (Kayed et al.,
tion, causes accumulation of a-synuclein (a-syn), 2003; Volles and Lansbury, 2003). However, the documentation
and results in neurotoxicity through aggregation- and characterization of these species in vivo have been
dependent mechanisms. Glucosylceramide (GlcCer), hampered by technical limitations as well as their evanescent
the GCase substrate, directly influenced amyloid nature in cells. Further, the in vivo factors that dictate the forma-
formation of purified a-syn by stabilizing soluble olig- tion and stabilization of these putatively toxic intermediates are
omeric intermediates. We further demonstrate that not known.
a-syn inhibits the lysosomal activity of normal GCase Recent description of a clinical link between GD and parkin-
in neurons and idiopathic PD brain, suggesting that sonism (Sidransky, 2005) suggested that mutations in the GCase
GCase depletion contributes to the pathogenesis of gene (GBA1) and alterations in sphingolipid metabolism
contribute to the pathogenesis of synucleinopathies. GD is
sporadic synucleinopathies. These findings suggest
a rare, autosomal recessive lysosomal storage disorder that
that the bidirectional effect of a-syn and GCase results from loss-of-function mutations in GCase, a lysosomal
forms a positive feedback loop that may lead to enzyme that cleaves the b-glucosyl linkage of GlcCer (Brady
a self-propagating disease. Therefore, improved tar- et al., 1965). Three types of GD have been described, based
geting of GCase to lysosomes may represent on the rate of clinical progression and involvement of the nervous
a specific therapeutic approach for PD and other system (Grabowski, 2008). Type I GD is classically defined as
synucleinopathies. non-neuronopathic and is typically characterized by hepatosple-
nomegaly, skeletal and hematopoietic system abnormalities.
Phenotypic variation in type I GD has been observed, and a small
INTRODUCTION subset of patients develop parkinsonism at variable ages
throughout the course of the disease (Bultron et al., 2010; Tayebi
The synucleinopathies, including dementia with Lewy bodies et al., 2003). Types II and III are differentiated from type I by neu-
(DLB) and Parkinson’s disease (PD), are a group of neurodegen- rodegeneration of the central nervous system with either rapid
erative disorders characterized by the accumulation of a-syn, (type II) or chronic progression (type III); however these forms
a small neural-specific protein involved in synaptic function can also show some phenotypic variation. A common feature
(Chandra et al., 2005). Although clinically diverse, the synuclei- of all three types is accumulation of GlcCer in the affected

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 37

tissues, but the reasons for phenotypic variability of GD are not in cell culture before the occurrence of more severe nuclear
known. toxicity (Zala et al., 2005). This revealed no change in neurotox-
The initial discovery of parkinsonism in a subset of adult onset icity when assessed at 7 days post-infection (dpi), suggesting
type I GD patients suggested a possible pathogenic link between that neurons have the ability to tolerate alterations in the GlcCer
the two disorders (Neudorfer et al., 1996; Sidransky, 2005; metabolizing pathway within this timeframe.
Tayebi et al., 2003). Neuropathological analysis of these patients We next analyzed proteolysis of long-lived proteins in living
revealed the presence of a-syn-positive Lewy bodies (Wong neurons and found that GCase KD significantly decreased the
et al., 2004), suggesting the involvement of a-syn aggregation. rate of proteolysis by 40% (Figure 1E, Figure S1B). To determine
It was subsequently noted that patients with GD and parkin- whether GCase KD affects a lysosomal degradation pathway,
sonism frequently had relatives with parkinsonism that were neurons were treated with the well-established lysosomal inhib-
heterozygous for GBA1 mutations (Goker-Alpan et al., 2004). itors ammonium chloride (NH4Cl) and leupeptin. These
Importantly, several additional genetic studies in large patient compounds did not additively inhibit the proteolysis in GCase
cohorts demonstrated that patients with parkinsonism have an shRNA-treated cells, indicating that GCase KD affects a lyso-
increased incidence of GBA1 mutations (Lill et al., 2008; Sidran- somal-mediated pathway (Figure 1E). Consistent with this,
sky et al., 2009), making GBA1 the most common known genetic immunofluorescence analysis of LAMP1, a lysosomal marker,
risk factor for PD to date. GBA1 mutations have also been iden- revealed accumulation and enlargement of LAMP1-positive
tified in patients with the diagnosis of DLB (Goker-Alpan et al., puncta in neurons (Figure S1G).
2006; Neumann et al., 2009), and inhibitors of GCase function Because Lewy bodies have been detected in postmortem
have been shown to modulate a-syn levels (Manning-Bog brain samples of patients with GD, we hypothesized that endog-
et al., 2009). While these studies provide correlative evidence enous a-syn protein may accumulate in neurons infected with
in patients that GlcCer metabolism may be linked to a-syn, the GCase shRNA. Indeed, GCase KD increased the steady-state
mechanism of such linkage has not been explored. Here, we levels of a-syn by 1.8-fold relative to controls, whereas the levels
show that intracellular GlcCer levels control the formation of of another disease-associated aggregation-prone protein, tau,
soluble toxic a-syn assemblies in cultured neurons and mouse did not change (Figure 1F). This also occurred without a change
and human brain, leading to neurodegeneration. The elevation in mRNA levels of a-syn, suggesting that the observed increase
and formation of a-syn assemblies further contributes to a path- in a-syn protein levels resulted from compromised protein
ogenic cycle by inhibiting the lysosomal maturation and activity degradation (Figure 1F). Analysis of a-syn levels after GCase
of normal GCase, resulting in additional GlcCer accumulation KD was also performed in a human neuroglioma cell line (H4) ex-
and augmented a-syn oligomer formation. pressing wild-type (WT) a-syn under the control of a tetracycline-
inducible promoter (‘‘tet-off’’). a-syn expression was turned off
RESULTS by dox to determine the a-syn degradation rate. This revealed
that GCase KD impeded the clearance of a-syn (Figure 1G).
Depletion of GCase Compromises Protein Degradation Taken together, these data suggest that KD of GCase in neurons
Capacity and Increases a-Syn Levels in Neurons leads to decreased lysosomal degradation capacity and conse-
The observation that loss-of-function GCase mutations cause quently increased levels of a-syn protein.
a-syn accumulation and Lewy bodies in GD brain prompted us To validate the primary culture results, dopaminergic neurons
to test the biological effects of GCase knockdown (KD) in were generated from induced pluripotent stem (iPS) cells derived
neurons. GCase shRNA-mediated KD by lentiviral infection re- from skin fibroblasts of a GD patient. Analysis of GD iPS cells re-
sulted in a 50% reduction in GCase protein levels compared to vealed the expression of Oct4, Tra-1-60, SSEA-4, and nanog,
nontransduced neurons control scrambled (scrb) shRNA-in- indicating that GD iPS cells contain the essential pluripotency
fected neurons (Figures 1A and 1B). The levels of mature lyso- factors, as well as normal chromosomal number, size, and
somal GCase were analyzed by endoglycosidase H (endo H) genomic structure (Figures S2A and S2B). Dopaminergic
treatment, an enzyme that only cleaves high mannose moieties neurons were induced from iPS cells by a previously established
of endoplasmic reticulum (ER) proteins. This revealed lower protocol (Seibler et al., 2011) to yield 80% of cells that ex-
levels of endo H-resistant GCase upon infection with GCase pressed the neuron-specific b III tubulin, and 10% that ex-
shRNA constructs, demonstrating a depletion of the lysosomal pressed the dopaminergic marker tyrosine hydroxylase (TH)
form (Figure S1A available online). Further analysis of whole- (Figure 2A). Genotyping analysis confirmed that GD, but not
cell lysates showed a decline in GCase activity (Figure 1B), WT, iPS neurons harbored the expected mutations in GCase
increased cellular lipids with BODIPY 493, and increased GlcCer (N370S/84GG insertion) and lower levels of GCase protein and
by immunofluorescence (Figures 1C and 1D). GlcCer accumula- activity (Figure 2B, Table S1). In addition, WT and GD cells did
tion was validated by mass spectrometry, which revealed a 4- not contain other mutations previously shown to cause PD (Table
fold increase of GlcCer in GCase-depleted neurons, whereas S1). Radioactive pulse-chase experiments in GD iPS neurons re-
the levels of ceramide and other sphingolipids remained vealed a decline in proteolysis of long-lived proteins compared
unchanged (Figure 1B, Figure S1D). Analysis of other lysosomal to WT cells, and the addition of lysosomal inhibitors did not
proteins and activity suggested that the constructs specifically further affect proteolysis (Figure 2C). Proteolysis measurement
decrease GCase protein (Figures S1C–S1F). Neurotoxicity was of short-lived proteins revealed no change compared to WT cells
assessed upon GCase KD by neurofilament (NF) immunostain- (Figure 2C, inset). Immunofluoresence and western blot analysis
ing, a sensitive method that detects the degeneration of neurites revealed a dramatic increase in a-syn protein levels in GD iPS

38 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.

Figure 1. GCase Knockdown Compromises Lysosomal Degradation and Causes Accumulation of a-Syn
(A) KD of GCase protein in cortical neurons by GCase shRNA is shown by western blot. Neural specific enolase (NSE) was used as a loading control. Four
replicates are shown. Scrb, scrambled shRNA.
(B) Left: GCase protein levels (n = 6, *p < 0.01). Middle: Enzymatic activity of GCase (n = 6, *p < 0.01). Right: Intracellular GlcCer quantification by MS (Pi,
phosphate) (n = 3, *p < 0.05).
(C) GlcCer immunofluorescence (top, red) and neutral lipids were visualized by BODIPY 493 fluorescence (bottom, green). Nuclei were visualized with DAPI (blue).
The arrows indicate cells with increased diffuse staining, whereas the arrowhead indicates a cell with punctated lipid accumulations.
(D) Fluorescent intensity shown in (C) was quantified and normalized to DAPI (n = 3, *p < 0.05).
(E) Proteolysis of long-lived proteins in neurons assessed at 8 hr. Lysosomal inhibitors leupeptin (leu) and ammonium chloride (NH4Cl) were used (n = 4, *p < 0.05).
(F) Western blot of endogenous a-syn (mAb syn202) and Tau. Four replicates are shown. Protein and mRNA levels are shown under the blots (n = 4, *p < 0.05).
a-Tub was used as a loading control.
(G) a-Syn analysis in inducible H4 cells. Expression was turned off by doxycycline (DOX) and protein clearance was measured by western blot with mAb syn211.
Quantifications are shown below (n = 6, *p < 0.05). GCase KD is shown by western blot and a-tub was used as a loading control. Molecular weight (MW) is
indicated in kDa. For all analyses, values are the mean ± standard error of the mean (SEM).
See also Figure S1.

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 39

Figure 2. Compromised Proteolysis of Long-Lived Proteins and Specific Accumulation of Endogenous a-Syn in Human GD Dopaminergic
Neurons
(A) Immunofluorescence analysis of WT and GD neurons generated from iPS cells with the neuronal marker b III tubulin (green) and catecholaminergic marker
tyrosine hydroxylase (TH, red). Nuclei (DAPI) are shown in blue. Scale bars = 10 mm.
(B) Western blot analysis of GCase. NSE was used as a loading control. Bottom, quantification of GCase activity (n = 3, *p < 0.05).
(C) Long-lived protein degradation was assessed as in Figure 1E (n = 4, *p < 0.05). Inset, proteolysis of short-lived proteins (15 min post-chase).
(D) a-Syn immunofluorescence analysis using mAb LB509 (red). b III tubulin, green. Scale bar = 30 mm.
(E) Western blot of T-sol lysates from iPS neurons. Htt, huntingtin; CBB, Coomassie brilliant blue.

40 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.

neurons compared to WT cells (Figures 2D and 2E). We did not between 14 and 39 kDa in size upon GCase KD (Figures 3D
observe changes in the levels of huntingtin and only mild and 3E). We next utilized native size exclusion chromatography
changes of tau in GD iPS neurons, indicating that GCase muta- (SEC) followed by SDS-PAGE/western blot of the collected frac-
tions primarily affect a-syn levels (Figures 2E and 2F). These data tions, to assess the presence of T-sol oligomeric a-syn species.
confirm that endogenous mutations in GCase affect lysosomal GCase KD resulted in the formation of high-molecular-weight
proteolysis and cause the preferential accumulation of a-syn. (HMW) assemblies with a molecular radius of 64–95 Å, in addition
to the normal monomeric form eluting as a 31–34 Å sized particle
Depletion of GCase Enhances a-Syn-Mediated (Figure 3F). Interestingly, analysis of D71–82 a-syn-expressing
Neurotoxicity through Aggregation-Dependent neurons revealed no change in the elution profile upon GCase
Mechanisms KD (Figure 3G), further indicating that GCase KD induces the
We next determined the effect of GCase KD on a-syn-mediated formation of a soluble HMW oligomeric a-syn that depends on
neurotoxicity. Because KD of GCase alone was not toxic at 7 dpi, the residues 71–82. These results further suggest that the ability
human a-syn was overexpressed by lentiviral transduction. Im- of a-syn to form soluble oligomers and insoluble species is a crit-
munostaining with human-specific anti-a-syn monoclonal anti- ical determinant for neurotoxicity induced by GCase KD.
bodies (mAbs) syn211 and LB509 revealed the expected punc- The increased a-syn levels and toxicity that occur with GCase
tate staining pattern in neuronal extensions consistent with depletion may result from generalized lysosomal inhibition or
a synaptic enrichment of a-syn (Figure S3A) (Maroteaux et al., may be due to alterations in GlcCer lipid metabolism. To distin-
1988). To examine the contribution of a-syn misfolding to neuro- guish between these two possibilities, we inhibited lysosomal
toxicity, we expressed the PD-linked A53T a-syn mutant as well protein degradation with leupeptin in WT a-syn-expressing
as an artificial fibrillization-incompetent mutant, D71-82 a-syn neurons and assessed neurotoxicity. We found that leupeptin
(Giasson et al., 2001), in primary neurons and observed treatment did not enhance a-syn-mediated neurotoxicity
increased levels of all three variants without neurotoxicity at (Figures S3D and S3H). Biochemical analysis revealed an
7 dpi (Figure 3 and Figure S3B). By contrast, expression of increase of T-insoluble a-syn in leupeptin-treated cells but no
human WT a-syn with GCase KD resulted in an 25% decline change in the amount of soluble a-syn (Figures 3D and 3E).
in viability by NF intensity and neuronal volume measurements This was corroborated by immunostaining analysis, which
compared to controls (Figures 3A and 3B). Western blot analysis showed an increase in the total a-syn immunostaining intensity
with mAb syn211 of Triton X-100 soluble (T-sol) lysates indicated in leupeptin-treated compared to control cells (Figure S3F).
that a-syn protein levels increased by 1.8-fold concomitantly SEC analysis also showed that soluble HMW a-syn were not
with the enhanced toxicity (Figure 3C). Importantly, KD of GCase detectable in neurons upon leupeptin treatment (Figure 3H),
also enhanced the toxicity of titer-matched A53T a-syn-infected consistent with their rapid consumption into insoluble species.
cells to the same extent as WT a-syn, whereas no toxicity was In addition, when comparing the increase of total a-syn (soluble
observed in D71–82 a-syn-expressing neurons (Figures 3A and and insoluble) by leupeptin treatment or GCase KD, we found
3B). Toxicity by WT a-syn expression/GCase KD was further that both approaches had similar effects (Figure S3F). Western
verified by measurement of condensed nuclei (Figure S3H, right). blot analysis also indicated a comparable increase in the levels
We also determined neuronal viability at later time points after of LC3-II, a well-established lysosomal substrate, by leupeptin
infection (10 dpi) and found that toxicity was further enhanced or GCase KD (Figure S3E). Thus, despite similar effects on the
in WT a-syn/GCase-depleted cells (50% viability assessed total a-syn levels by leupeptin or GCase KD, only GCase KD
by NF intensity) (Figure S3C). Because GCase KD resulted in increased the steady-state levels of soluble HMW a-syn. Also,
increased levels of A53T and D71–82 a-syn proteins to a similar we used sequential extraction followed by SDS-PAGE/Coomas-
extent as WT a-syn (Figure 3C), the toxicity appears to depend sie brilliant blue (CBB) staining to determine the effect of leupep-
on amino acids 71–82 of a-syn, a mostly hydrophobic region tin treatment on the solubility of total cellular proteins. Interest-
that is required for a-syn polymerization (Giasson et al., 2001). ingly, we found that whereas leupeptin treatment increased the
Taken together, these results suggest that GCase KD promotes levels of total insoluble proteins by 2-fold, GCase KD had no
the accumulation and neurotoxicity of a-syn through polymeriza- effect (Figure S3G). This indicates that GCase KD preferentially
tion-dependent mechanisms. affects the solubility of a-syn. Taken together, these data
suggest that alterations in the GlcCer metabolic pathway influ-
Enhanced a-Syn-Mediated Neurotoxicity by GCase ence the formation of toxic soluble and insoluble a-syn species,
Depletion Is Dependent on the Formation of Soluble and causing a stabilization of soluble HMW forms of the protein.
Insoluble High-Molecular-Weight Species of a-Syn
To directly determine whether GCase KD affects a-syn polymer- GlcCer Influences the Aggregation of a-Syn In Vitro
ization in neurons, lysates were sequentially extracted and sepa- by Stabilizing Soluble Oligomeric Intermediates
rated into T-sol and -insoluble fractions followed by western blot Because GCase KD significantly affected a-syn aggregation in
with mAb LB509. This revealed an increase of T-sol monomeric neurons, we next examined whether GlcCer directly influences
a-syn (18 kDa), as well as T-insoluble a-syn species migrating the in vitro aggregation of recombinant a-syn. Lipid dispersions

(F) Quantification of a-syn, tau, and Htt protein by western blot. Protein levels were normalized to a-tub (n = 3, values are the mean ± SEM, ***p < 0.01 compared to
WT a-syn, WT and GD tau, and WT and GD Htt; *p < 0.05 compared to WT tau).
For all quantifications, values are the mean ± SEM. See also Figure S2 and Table S1.

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 41

Figure 3. GCase Depletion Enhances a-Syn-Mediated Neurotoxicity through Aggregation-Dependent Mechanisms
Neurons expressing human a-syn proteins and GCase shRNA were analyzed at 7 dpi.
(A) Neurofilament immunostaining was used to monitor neurite degeneration. Representative neurofilament immunofluorescence staining (green) in WT a-syn-
expressing neurons is shown below. Nuclei (DAPI) are shown in blue. Scale bars = 10 mm.
(B) Neurotoxicity was assessed by neuronal volume analysis. (for A and B: n = 8, *p < 0.001.)
(C) Protein levels of human WT, A53T, and D71–82 a-syn (T-sol) by western blot. a-tub was used as a loading control. Quantification is shown below (n = 6,
*p < 0.01).
(D) a-Syn western blot of T-sol fractions (leu, leupeptin; NT, not transduced). NSE was used as a loading control.
(E) Western blot of T-insoluble a-syn. Quantification is shown below. The brackets show the signal used for quantification (n = 3, *p < 0.05, **p < 0.01 compared to
scrb control).
(F–H) Native SEC/western blot analysis of T-sol lysates (Å, radius in angstroms). NSE was used as a loading control. Oligomeric a-syn (Void / 64 Å) was
quantified (fold change: scrb shRNA = 1 ± 0.5; GC shRNA = 19.5 ± 6.0) (n = 3, values are the mean ± SEM, *p < 0.05).
MW is indicated in kDa for each blot. For all quantifications, values are the mean ± SEM. See also Figure S3.

42 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.

Figure 4. GlcCer Directly Influences the In Vitro Fibril Formation of Recombinant a-Syn and Stabilizes Soluble Oligomeric Species
(A) Purified a-syn was incubated with mixtures of PC and GlcCer at pH 5.0, 37 C and amyloid formation was assessed by thioflavin T fluorescence (relative
fluorescence units [RFU], n = 4, *p < 0.01).
(B) Analysis of 100,000 3 g soluble a-syn at 1 and 5 hr by SEC (115–38 Å and 36–27 Å fractions), then SDS-PAGE/western blot (syn211). The MW is indicated in
kDa.
(C) Soluble oligomers were quantified by densitometry (n = 3, *p < 0.05).
(D) ANS fluorescence of a-syn species formed after 1 hr (n = 4, *p < 0.01).
(E) Centrifugal sedimentation analysis at 28 hr (s, supernatant; p, pellet). a-Syn was detected with Coomassie brilliant blue staining. Pelletable a-syn was
quantified in the graph below (n = 3). Amyloid was measured from the same reactions by thioflavin T (n = 4, *p < 0.01).
(F) EM analysis of a-syn aggregates showing a mixture of fibrillar (i–iii) and amorphous (iv–v) structures at 24 hr. Panels ii–v show immmuno-EM analysis using
mAb syn505. Scale bars: 100 nm for i–iii; 500 nm for iv and v.
(G) Immuno-EM analysis with syn505 of a-syn+PC25/GlcCer75 reactions at 15 hr. GlcCer lipid tubules are 50 nm in width. Scale bars: 100 nm for i and iii; 500 nm
for ii.
(H) Immuno-EM analysis with syn505 of a-syn+PC25/GlcCer75 reactions at 24 hr showing fibrillar structures of 10–14 nm in width with twisted (i) or straight (ii)
morphologies that appear to extend from GlcCer tubules. Scale bars: 100 nm.
(I) Immuno-EM analysis of GlcCer lipid dispersions alone. Scale bar: 100 nm.
For each graph in (A) and (C)–(E), values are the mean ± SEM. See also Figure S4.

made of mixtures of purified GlcCer and brain phosphatidylcho- aggregation under physiological conditions showed that GlcCer
lines (PCs) were incubated with a-syn at physiological conditions had no effect on fibril formation (Figure S4), consistent with
(pH 7.4, 37 C). Electron microscopy (EM) analysis indicated the previous observations (Martinez et al., 2007).
formation of tubules consisting of polymerized GlcCer (Figures We next assessed the effect of GlcCer on a-syn fibril formation
4G–4I), similar to those previously observed in Gaucher cells in under acidic conditions (pH 5.0, 37 C) to simulate a lysosome-
patients and mouse models (Lee, 1968). The analysis of a-syn like environment in vitro because our neuronal culture data

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 43

indicated increased colocalization of a-syn with LAMP1 upon accelerated fibril assembly once this phase was initiated (Fig-
GCase KD (Figures S3H and S3I, Extended Discussion). These ure 4A). The fibril assembly phase of PC25/GlcCer75-containing
experiments revealed that acidic reactions containing lipid reactions occurred between 16 and 24 hr, compared to control
dispersions made of 90% PC and 10% GlcCer (PC90/GlcCer10) reactions where the assembly occurred between 2 and 24 hr.
did not significantly influence the fibril formation of a-syn Furthermore, the maximal thioT signal at the end stages of the
compared to control reactions containing a-syn alone (Figure 4A, reaction was 2- to 3-fold higher compared to control reactions
Figure S4A). However, increasing the amount of GlcCer to 75% (Figure 4A). We further analyzed the aggregated species formed
while keeping the total lipid amount constant (PC25/GlcCer75) at the end stage of the fibril-forming reaction, after assembly was
altered the kinetic profile of a-syn fibril formation by delaying completed and equilibrium was reached (at 28 hr). Centrifugal
the formation of insoluble thioT-positive a-syn fibrils, extending sedimentation analysis at 100,000 3 g, which detects both
the lag time from 2 to 16 hr (Figure 4A). amyloid and non-amyloid aggregates in the pelletable (P) frac-
Because our biochemical data from cell culture experiments tions, revealed that GlcCer had no effect on the amount of pellet-
suggested that GlcCer selectively increased soluble HMW forms able a-syn protein (Figure 4E). In the same reaction mixtures
of a-syn (Figure 3), we hypothesized that the delay in fibril forma- used for sedimentation analysis, measurement of amyloidogenic
tion observed in vitro resulted from a kinetic stabilization of a-syn with thioT revealed a 3-fold increase in the amount of
a soluble oligomeric intermediate species. To test this, the nature amyloid detected in PC25/GlcCer75-containing reactions (Fig-
of the species that form during the lag phase (between 1 and ure 4E, bottom). Immuno-EM analysis of a-syn/GlcCer reactions
16 hr) of PC25/GlcCer75-containing reactions was characterized at 24 hr confirmed the presence of 14 nm wide fibrillar struc-
by analytic biochemical methods. Soluble portions of the reac- tures that appeared to extend from GlcCer tubules (Figure 4H),
tion mixtures were obtained by centrifugation at 100,000 3 g whereas a-syn-alone reactions contained both fibrillar (Figure 4F,
and analyzed at 1 and 5 hr after the addition of lipids by SEC/ i–iii) as well as amorphous aggregates (Figure 4F, iv and v). Taken
SDS-PAGE. This revealed an increase in the amount of HMW together, these data indicate that GlcCer selectively stabilizes
oligomeric a-syn eluting between 115 and 38 Å, and migrating the formation of soluble oligomeric intermediates on-pathway
at 18 kDa by SDS-PAGE, in samples containing PC25/GlcCer75 to forming amyloid fibrils. However, due to the continuous accu-
lipid dispersions (Figure 4B). Further, we detected increased mulation of these soluble on-pathway intermediates that occurs
amounts of soluble SDS and heat-stable dimers (36 kDa), trimers in vitro between 2 and 16 hr, the concentration of GlcCer is likely
(54 kDa), and higher oligomeric species eluting as 36–27 Å-sized eventually surpassed and results in the rapid formation of thioT-
particles in PC25/GlcCer75-containing reactions compared to positive amyloid fibrils.
controls (Figure 4B). The GlcCer-induced soluble oligomeric
species appeared to increase between 1 and 5 hr, whereas olig- Accumulation of Soluble and Insoluble a-Syn Species
omers and monomers in control reactions decreased, consistent Occurs in GD Mouse
with their consumption into insoluble fibrils (Figures 4B and 4C). As GlcCer appears to affect the levels of soluble a-syn oligo-
Native gel electrophoresis also revealed an increase in the mers in neuronal cultures and in vitro, we next examined
amount of 720–1048 kDa-sized a-syn species (Figures S4C whether GCase depletion and GlcCer accumulation affect
and S4D). Further, we found that other sphingolipids did not a-syn levels and soluble oligomer formation in vivo. For this,
significantly alter the amounts of soluble oligomers, indicating brain tissues from a previously described GD mouse model
a specific effect by GlcCer (Figures S4E and S4F). Immuno-EM (4L/PS-NA) were analyzed to determine whether endogenously
with syn505 antibodies that preferentially detect misfolded expressed a-syn protein levels were elevated. Previous analysis
a-syn demonstrated the formation of individual spherical struc- of this mouse model indicated low levels of GCase activity,
tures of 25–50 nm in diameter that occasionally appeared to neuronal accumulation of GlcCer, and severe neurological dete-
coalesce to form larger amorphous structures (Figure 4G, iii). rioration by 20 weeks of age (Sun et al., 2005). In addition to
Syn505 also detected a-syn directly on GlcCer tubular structures GlcCer, the levels of other sphingolipids were also determined
(Figure 4G, i and ii) but not on GlcCer-alone reactions (Figure 4I), showing an accumulation of lactosylceramide, GM2, and
indicating an association of misfolded a-syn with GlcCer. a-Syn- GD3, whereas ceramide levels remained unchanged (Figures
GlcCer reactions were further analyzed by 8-anilino-1-naptha- S5A and S5B). Our neuropathological analysis revealed the
lene sulfonate (ANS) binding, a fluorescent dye used to detect presence of eosinophilic spheroids, suggesting the presence
aggregation-prone conformational intermediates (Stryer, 1965). of degenerating neurons, in multiple brain regions including
Enhanced ANS fluorescence was observed in soluble a-syn the substantia nigra (SN) and cortex (Ctx) in GD mice compared
samples incubated with PC25/GlcCer75 compared to control to WT mice that exhibited normal neuronal architecture (Figures
reactions, indicating that GlcCer addition results in a conforma- 5A and 5D). These degenerative changes occurred concomi-
tional alteration that increases solvent-exposed hydrophobic tantly with increased levels of a-syn in these regions (Figure 5B).
regions (Figure 4D). Because hydrophobicity changes in proteins Immunofluorescence analysis revealed the presence of a-syn
correlate with aggregation propensity, this observation indicates accumulations in the form of punctated structures (<5 mm in
that GlcCer stabilizes the formation of a soluble assembly- diameter), whereas WT mice showed a normal neuropil staining
competent intermediate species during the lag phase of the fibril pattern expected for a-syn (Figures 5B–5D). These changes
formation reaction. were not restricted to the SN and Ctx, as a-syn accumulations
Further inspection of the kinetic profile indicated that although were also observed in other neural regions including cere-
GlcCer delayed the onset of fibril formation from 2 to 16 hr, it also bellum, hippocampus, and brainstem (Xu et al., 2010).

44 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.

Additionally, both intraneuronal and extraneuronal a-syn accu- at 42–45 Å (or 110–140 kDa globular protein). The a-syn in
mulations were identified by costaining with the neuron-specific this oligomeric fraction, when analyzed by denaturing SDS-
marker NeuN (Figure 5C), whereas quantitative analysis did not PAGE, resolved as 22, 44, and 75 kDa heat-stable species (Fig-
reveal significant neuronal loss (Figure 5D). The solubility of ure 6F). These data demonstrate that elevated T-sol a-syn in
a-syn was analyzed in 4L/PS-NA by sequential extraction in the form of oligomeric species is present primarily in the GD/
Triton X-100 buffer, then 2% SDS buffer. Both syn202 and APD brain.
SNL-1, antibodies that detect total a-syn, revealed increased We also detected elevated levels of a-syn oligomers in
levels of T-sol a-syn in 4L/PS-NA mice compared to WT mice patients that were homozygous or heterozygous for GCase
(Figures 5E, left, and 5F). T-insoluble fractions showed the ex- mutations (Table S2) with a diagnosis of Lewy body dementia
pected low levels of a-syn in WT mice and more aggregated (DLB) (Figures 6G and 6K). Analysis of postmortem brain lysate
a-syn in 4L/PS-NA mice as detected with both syn202 and obtained from infants diagnosed with type II GD as well as a 3-
syn505 (Figures 5E, right, and 5F). Analysis of T-sol levels by year-old child diagnosed with type III GD also exhibited
SEC showed increased levels of putative oligomeric forms increased oligomeric a-syn eluting above 36 Å (Figures 6H–6J),
(120–70 Å- and 51–44 Å-sized species), whereas monomers although some variation between samples was observed. We
were similar to control mice (Figures 5G and 5H). Quantification quantified the levels of oligomeric a-syn detected with the
of the soluble HMW a-syn revealed a 4-fold increase in 4L/PS- syn211 mAb and found that both homozygote and heterozygote
NA mice compared to control mice (Figures 5F and 5H). We carriers of GBA1 mutations with a neuronopathic phenotype
confirmed our analysis of a-syn in another previously described contained significantly higher levels of oligomers compared to
and well-characterized GD mouse model, GCase harboring the controls (Figure S6C). We also verified that these GD samples
GD-linked D409H loss-of-function mutation (Xu et al., 2003). contained lower GCase protein and activity levels (Figures
This revealed that D409H mice had similar increases in a-syn S6A, S6B, and S6E). Taken together, the data suggest that
punctated structures as observed by immunostaining analysis GCase deficiency promotes the formation of oligomeric a-syn,
(Figure S5C) as well as higher levels of soluble oligomers and and the occurrence of these oligomers in type II and type III
insoluble a-syn species (Figure S5D). Finally, we used a well- GD brain suggests that they may also play a role in the pathogen-
established C. elegans model to further demonstrate that deple- esis of age-independent, infantile GD forms.
tion of GCase causes the accumulation of a-syn in vivo We further analyzed the 45 Å-sized oligomeric fractions with
(Figure S5E). Taken together, these data are consistent with mAb syn303, an antibody that preferentially detects pathological
our cell culture and in vitro data demonstrating that GCase oligomeric a-syn (Duda et al., 2002) and can distinguish poten-
depletion promotes the formation of soluble oligomeric and tially toxic from nontoxic a-syn species (Tsika et al., 2010). We
insoluble a-syn in vivo. found that syn303 immunoreactivity was increased in all of the
neuronopathic GD samples (Figure 6L, Figure S6D). In most of
Elevated Levels of Soluble HMW a-Syn in GD Brain Are the cases, syn303 reacted with the 22, 44, and 75 kDa species
Associated with Neurodegeneration that were also detected with syn211 (Figure 6L). These data
Because our in vitro, cell culture, and GD animal model data further demonstrate that toxic oligomeric a-syn is elevated in
suggested that GlcCer accumulation leads to elevated levels patients harboring GBA1 mutations and is preferentially associ-
of soluble a-syn oligomers, we next sought to identify these ated with neuronopathic forms of the disease.
species in human postmortem brain samples obtained from
patients with GD. T-sol fractions of cortical samples were Overexpression of a-Syn Inhibits the Intracellular
analyzed by native SEC, followed by SDS-PAGE/western blot Trafficking of GCase Resulting in Decreased Lysosomal
of the collected fractions using mAb syn211. Analysis of healthy GCase Activity
controls without GBA1 mutations (Table S2) revealed the ex- Although most patients with idiopathic PD invariably have
pected elution profile typically observed for monomeric human elevated levels of a-syn protein, they do not harbor mutations
a-syn, eluting mainly as a 34 Å-sized particle by SEC and in the GBA1 gene and thus are expected to have normal lyso-
migrating at 18 kDa by SDS-PAGE (Figures 6A–6C). Analysis somal function of GCase. However, recent evidence in
of cortical T-sol lysate from two pathologically and clinically S. cerevisae and cell lines indicated that overexpression of
confirmed non-neuronopathic type I GD patients revealed an a-syn has the ability to impede ER–Golgi trafficking of proteins,
a-syn elution profile that was similar to control (Figures 6D by inhibiting the formation of soluble N-ethylmaleimide-sensi-
and 6E), although the total levels of monomeric a-syn were tive factor attachment protein receptor (SNARE) protein
elevated (a-syn protein levels, % of control): control = 100 ± complexes (Cooper et al., 2006; Thayanidhi et al., 2010). To
12.6, GD type I (no PD) = *243 ± 53,values are the mean ± stan- determine whether a-syn disrupts lysosomal maturation and
dard error of the mean (SEM),*p < 0.05, n = 3 controls, n = 2 GD activity of GCase, we overexpressed a-syn in H4 cells and
type I). When brain lysate from a previously documented GD primary cortical neurons that express WT GCase and measured
patient diagnosed with atypical Parkinson’s disease (APD) the post-ER forms. The intracellular trafficking of GCase was
was analyzed (Tayebi et al., 2001), a dramatic increase in assessed by SDS-PAGE/western blot, through molecular
a-syn levels was observed (Figure 6F). a-Syn eluted as weight (MW) analysis of various GCase forms that result from
a 34 Å-sized particle and migrated at 18 kDa by SDS-PAGE protein glycosylation. Whereas the ER form of GCase migrates
similar to controls, but a substantial proportion (50% of the at 60 kDa, the post-ER GCase forms migrate above 60 kDa
total T-sol) also eluted as a larger putative oligomeric species (Erickson et al., 1985). Analysis of whole-cell lysates from

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 45

46 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.
inducible H4 cells showed that lowering a-syn expression GBA1 mutations and between the ages of 65 and 80 years of
levels by the addition of dox for 24 or 32 hr resulted in age (Table S3 and Table S4), we noticed a natural variability of
a concomitant increase in the post-ER GCase forms while a-syn expression levels between subjects. Control samples
decreasing the 60 kDa ER form (Figure 7A). Similarly, overex- 1, 2, 4, and 6 were noted to have mid-to-high levels of a-syn rela-
pression of human WT and A53T a-syn in primary cortical tive to samples 3 and 5, which contained very low a-syn levels
neurons also altered the post-ER/ER GCase ratio by causing (Figure 7D). When the GCase glycosylation patterns were
an accumulation of the ER form, as well as a decrease in the analyzed by western blot, we observed a dramatic difference
post-ER forms migrating above 60 kDa (Figure 7B). Titer- in the post-ER/ER GCase ratio that correlated with a-syn levels.
matched infection of WT and A53T a-syn containing plasmids While all samples appeared to have similar levels of post-ER
resulted in almost equal alterations in the post-ER/ER GCase GCase, samples with low a-syn (samples 3 and 5) contained
ratio, despite the lower protein expression of A53T, indicating much less of the 60 kDa ER form (Figure 7D). Endo H digestion
that A53T more potently inhibits GCase trafficking compared also confirmed higher levels of ER-containing GCase, which
to the WT protein (Figure 7B). Interestingly, expression of migrated below 60 kDa after endo H treatment (Figure S7F).
D71–82 a-syn at levels that were slightly higher than WT We further analyzed the GCase activity levels in cortical tissue
a-syn caused only a mild alteration in the post-ER/ER GCase from whole-cell lysates of all low and high a-syn-containing
ratio that was not significantly different compared to empty samples and found no difference in activity (Figure S7H, left).
vector controls (Figure 7B). To verify that alterations in GCase However, when P2 and P3 GCase activity was determined, we
glycoslyation patterns observed by western blot corresponded found that microsome-enriched P3 fractions of ‘‘high’’ a-syn
to lower lysosomal activity of GCase, enzymatic activity was samples contained significantly higher levels of activity whereas
measured in lysosomal (P2) and microsomal (P3) enriched frac- no change was observed in the P2 fraction (Figure S7H). Western
tions (Figures S7A and S7B) of primary neuronal cultures. In P2 blot analysis with syn303 also revealed higher levels of oligo-
fractions, expression of both WT and A53T a-syn resulted in meric, oxidized a-syn in ‘‘high’’ a-syn sample C6 compared to
a significant decrease in GCase activity and a concomitant C5 (Figure S7G). These findings suggest that normal variation
increase in the P3 activity compared to controls (Figure 7C). of a-syn protein levels modulate the lysosomal maturation and
Conversely, the expression of D71–82 a-syn did not affect activity GCase in vivo.
GCase lysosomal activity (Figure 7C). The results were validated The observation that elevated a-syn levels affect GCase traf-
by endo H treatment of lysates, which revealed higher levels of ficking in neurons led us to hypothesize that lysosomal GCase
endo H-sensitive GCase migrating below 60 kDa in endo function may be lowered in idiopathic PD brain. We found an
H-treated samples of WT a-syn-expressing cells compared to 40% decline in the total GCase protein levels in T-sol lysates
control conditions (Figure S7C). Additionally, we tested the from cingulate cortex of PD brain when compared to age- and
ability of another amyloid-forming protein, poly Q-expanded postmortem time-matched controls (Figure 7E). In addition,
huntingtin fragment 548–72Q, to inhibit GCase maturation and there was an 50% decline in GCase activity in the P2 fraction
found no change (Figure S7C). Finally, we confirmed that the relative to age-matched controls, whereas no change was
accumulation of ER GCase by a-syn occurred at the protein observed in the P3 fraction (Figure 7E, bottom). Genotyping anal-
level, as measurement of GBA1 mRNA of a-syn-expressing ysis revealed that these patients did not harbor mutations in
neurons was not different compared to control conditions (Fig- GBA1, with the exception of one sample that contained the
ure S7D). The enzymatic activity of other lysosomal proteins in heterozygous mutation N370S (Table S5). This one sample,
the P2 fraction of infected neurons revealed only minor alter- PD4, had lower than the expected 50% decline in GCase activity
ations by a-syn expression, suggesting a preferential effect of (38% of control), possibly indicating additional inhibition of
a-syn on GCase activity (Figure S7E). GCase by a-syn accumulation (Table S6). Taken together, these
To determine whether GCase glycosylation patterns are sensi- data suggest that elevated levels of a-syn observed in PD and
tive to a-syn protein levels in vivo, human cortical material was other synucleinopathies lead to decreased lysosomal activity
analyzed by GCase western blot. Upon analyzing brain tissue of normal GCase that may in turn contribute to further propaga-
from several reportedly healthy controls without common tion and stabilization of oligomeric a-syn.

Figure 5. a-Syn Accumulation and Soluble Oligomer Formation in GD Mice

Analysis of 12-week-old GD mice (4L/PS-NA).
(A) H & E stain of the substantia nigra (SN) and cortex (Ctx). The arrows indicate eosinophilic spheroids. Scale bars = 50 mm.
(B) Immunofluorescence of a-syn (red) in SN and Ctx. Nuclei are stained with DAPI (blue). Scale bars = 20 mm.
(C) Costaining of a-syn (red) and neuronal marker NeuN (green). Scale bars = 20 mm.
(D) Left: Quantification of neuronal spheroids. ND, not detected. Middle: Quantification of neuronal number by NeuN immunostaining. Right: Quantification of
a-syn aggregates by immunostaining.
(E) Sequential extraction analysis of Ctx. pAb SNL-1 and mAb syn202 detect total endogenous a-syn, whereas syn505 detects oxidized/nitrated and misfolded
a-syn. NSE and a-tub were used as loading controls.
(F) Quantification of T-sol monomers (18 kDa, left), T-sol oligomers (>18 kDa, middle), and T-insoluble a-syn (total lane, right).
(G) Native SEC/SDS-PAGE/western blot of T-sol fractions. Radius, Å.
(H) Chromatographic profile obtained by syn202 densitometry. The values are representative of independent SEC analyses from three mice. The MW is indicated
in kDa for each blot. For all quantifications, values are the mean ± SEM.
See also Figure S5.

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 47

Figure 6. Accumulation of T-Sol a-Syn Oligomers Occurs in GD Brain
Native SEC followed by SDS-PAGE/western blot of human cortical lysates (T-sol). Radius is in Å (horizontal), apparent MW is in kDa (vertical). Monomeric a-syn
elutes at 34 Å.
(A–C) Healthy controls.
(D and E) Type I non-neuronopathic GD.
(F) Atypical Parkinson’s disease (APD).
(G) dementia with Lewy bodies (DLB).

48 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.

DISCUSSION ations in the GlcCer metabolism pathway influence a-syn oligo-
merization that is not necessarily age dependent.
Our data indicate that deficient GCase leads to accumulation of The absence of oligomeric a-syn in samples from type I GD
GlcCer in neurons that in turn promotes formation of toxic without parkinsonism (Figures 6D and 6E) indicates that other
a-syn oligomers. GCase depletion causes a decline in lysosomal factors, in addition to deficiency of GCase, likely contribute to
proteolysis that preferentially affects a-syn (Figures 1E and 1F). It oligomerization of a-syn in neuronopathic GD. For example,
is likely that the unique inherent property of a-syn to form amyloid oxidation and nitration of a-syn have been shown to impede
fibrils plays a critical role in the neurotoxicity that occurs with clearance and stabilize a-syn oligomers in vitro (Hodara et al.,
GCase KD, as expression of a-syn mutant lacking the 71–82 2004), and chaperones have also been shown to abrogate
region did not affect neuronal viability in our culture model (Fig- a-syn toxicity and aggregation (Auluck et al., 2002). Although
ure 3). Interestingly, another aggregation-prone protein involved our analysis of GD brain indicated increased levels of oxidized
in neurodegenerative disorders, tau, did not accumulate (Figure 1 a-syn in neuronopathic forms (Figure 6L), further studies are
and Figure 2), indicating that GCase function is preferentially required to examine how oxidation and other age-dependent
related to a-syn. Importantly, similar results were observed in processes interact with deficiency of GCase in promoting
human iPS neurons in the presence of endogenous mutations a-syn oligomerization.
in GCase, suggesting that GCase protein depletion or expression Our data also demonstrate that elevated a-syn inhibits intra-
of loss-of-function GCase mutants exhibit comparable cellular trafficking and lysosomal function of normal GCase in
phenotypes. neurons (Figure 7 and Figure S7). This indicates that decreased
Biochemical analysis indicated that GCase KD caused GCase activity not only contributes to toxicity in patients with
a dramatic increase in the levels of soluble oligomers (Figure 3). GBA1 mutations but also may affect the development of more
This effect was distinct from that observed by leupeptin, which common sporadic forms of PD and synucleinopathies that do
primarily resulted in elevated levels of T-insoluble a-syn not have mutations in the GBA1 gene. Interestingly, we show
without altering soluble forms. This suggests that in addition that variation of a-syn levels in healthy control subjects can
to general lysosomal inhibition, GlcCer accumulation specifi- also alter ER-Golgi flux of GCase, a property that may be poten-
cally affects the conformation and solubility of a-syn by stabi- tiated by a-syn oligomerization. This is further suggested by
lizing the levels of soluble intermediates. In vitro studies using normal GC activity in neurons expressing aggregation-incompe-
purified recombinant a-syn demonstrated that GlcCer has the tent D71–82-a-syn (Figure 7C) as well as the increased immuno-
ability to prolong the lag phase of fibril growth and stabilize reactivity to syn303 in controls that contain higher levels of ER
oligomeric intermediates only at acidic pH (Figure 4 and Fig- GCase (Figure S7G). However, further studies are required to
ure S4). This pH-dependent effect is consistent with accumula- delineate the precise mechanism of a-syn-mediated inhibition
tion of a-syn within LAMP1-positive vesicles and subcellular of GCase maturation.
fractions upon GCase KD (Figures S3H and S3I). After the Taken together, our results suggest that elevated levels of
lag phase, GlcCer accelerated amyloid formation and formed toxic a-syn species lead to depletion of lysosomal GCase and
fibrils that appeared to extend from GlcCer lipid tubules (Fig- further stabilization of a-syn oligomers by GlcCer accumulation.
ure 4H). It is possible that GlcCer tubules provide a scaffold This self-propagating positive feedback process proceeds until
or platform for oligomeric intermediates to form that, once a pathogenic threshold is surpassed, resulting in neurodegener-
saturated, proceed to rapid polymerization of fibrils. This ability ation (Figure 7F). Therefore, specific treatments that promote
may be a crucial step in pathogenesis, as the documentation targeting of GCase to lysosomes are expected to diminish the
of a-syn oligomers appears to be correlated with neurodegen- formation of toxic a-syn oligomers and break the pathogenic
eration in neuronal cultures, mouse models, and human neuro- cycle of a-syn aggregation and toxicity in PD and other
nopathic GD brain. synucleinopathies.
SEC analysis of postmortem GD and PD brain demonstrated
elevated levels of a previously undocumented 36–45 Å-sized EXPERIMENTAL PROCEDURES
soluble oligomeric a-syn species that correlated with a neurolog-
ical phenotype. The oligomers prominently reacted with the mAb Primary Cortical Cultures, Lentiviral Infection, and Leupeptin
syn303, an antibody generated against oxidized/nitrated a-syn Treatment
that preferentially detects pathological conformations of the Primary cortical culture procedures have been described in detail previously
(Tsika et al., 2010). Cells were infected at a multiplicity of infection (moi) of 3
protein that exhibit toxic properties (Tsika et al., 2010). The path-
for both GCase shRNA and a-syn-expressing lentivirus. For leupeptin treat-
ological a-syn oligomers were also detected in infantile neurono- ment, cells were infected with a-syn-expressing lentivirus at days in vitro
pathic GD cases, and in a child with type III GD (Figure 6), (DIV) 5, then treated with 50 mM leupeptin (EMD chemicals, http://www.
strongly suggesting that GBA1 mutations and specific alter- emdchemicals.com) at DIV 8, and harvested at DIV 12 (or dpi 7).

(H and I) Analysis of cortical material obtained from infants with type II acute neuronopathic GD.
(J) Cortical lysates from a 3-year-old child with neuronopathic type III GD.
(K) DLB with a heterozygous mutation in GBA1.
(L) Analysis of the 45 Å-sized fraction with syn303, which preferentially detects pathological oligomeric a-syn. Bands migrating at 18, 44, and 75 kDa were
detected with both syn303 and syn211 (arrows).
See also Figure S6 and Table S2.

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 49

Figure 7. Elevated Levels of a-Syn Inhibit the Intracellular Trafficking of GCase and Decrease Lysosomal GCase Function
(A) Inducible H4 cells expressing human WT a-syn were analyzed by western blot for post-ER and ER GCase (n = 6, *p < 0.01). a-tub was used as a loading control.
(B) post-ER/ER GCase in cortical neurons expressing human WT, A53T, or D71–82 a-syn. a-syn levels were determined by syn211 (human-specific) and syn202
(human and mouse). NSE was used as a loading control.
(C) GCase activity in cortical neurons of P2 and P3 fractions (n = 6, *p < 0.01, compared to vect).
(D) Analysis of GCase in cortex of 65- to 80-year-old controls. Samples 1, 2, 4, 6 = ‘‘high a-syn’’; samples 3, 5 = ‘‘low a-syn.’’ Quantification of a-syn protein and
post-ER/ER GCase levels is graphed below the blots (*p < 0.01).
(E) GCase western blot of PD brain lysates. a-Tub and CBB were used as loading controls. GCase levels were quantified below (n = 3 [control] or 6 [PD], *p = 0.02).
Bottom: GCase activity in P2 and P3 fractions (n = 3–6, *p = 0.04). MW for each blot is indicated in kDa.
(F) Pathogenic positive feedback mechanism of a-syn and GCase depletion in the lysosome. (1) Lysosomal GlcCer accumulation accelerates and stabilizes soluble
a-syn oligomers (bold arrow), which eventually convert into amyloid fibrils (thin arrow). (2) Accumulation of a-syn blocks the ER-Golgi trafficking of GCase. (3) Decrease
of GCase in the lysosome further amplifies GlcCer accumulation and stabilization of soluble a-syn oligomers and results in a stronger inhibition of GCase ER-Golgi
trafficking with each pathogenic cycle. For all quantifications, values are the mean ± SEM. See also Figure S7 and Table S3, Table S4, Table S5, and Table S6.

50 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.

Proteolysis Measurements ACKNOWLEDGMENTS
Proteolysis of long-lived proteins was determined by radioactive pulse-chase
using H3-leucine (Extended Experimental Procedures). We thank Harry Ischiropoulos for pCDNA3.1 human a-syn plasmids and
syn303 antibody, Benoit I. Giasson for the SNL-1 antibody, and Kimberly Kegel
Neurotoxicity Assessment for technical advice on liposome formation. This work was supported by
Cortical cells were seeded in 96-well plates at 50,000 cells/well, infected at DIV National Institutes of Health grants R01NS051303 (D.K.) and F32NS066730
5, and fixed in 4% paraformaldehyde at the indicated time points. The staining (J.R.M.) from the National Institute of Neurological Disorders and Stroke,
and analysis procedures have been described (Tsika et al., 2010). R01DK36729 (G.A.G.), and the Intramural Programs of the National Human
Genome Research Institute and the National Institutes of Health (E.S.).

Sequential Biochemical Extraction of Cell Cultures and Tissues

Received: November 4, 2010
Cells were harvested in Triton X-100 lysis buffer. The extracts were centri-
Revised: February 23, 2011
fuged at 100,000 3 g for 30 min. The pellets were extracted in 2% SDS
Accepted: May 27, 2011
buffer. Similar procedures were utilized for mouse and human brain
Published online: June 23, 2011
tissues, using 20 volumes of Triton X-100 lysis buffer. Samples were
loaded onto SDS-PAGE gels or subjected to native SEC followed by
REFERENCES
western blot analysis (Extended Experimental Procedures) (Mazzulli et al.,
2006).
Auluck, P.K., Chan, H.Y., Trojanowski, J.Q., Lee, V.M., and Bonini, N.M.
(2002). Chaperone suppression of alpha-synuclein toxicity in a Drosophila
Native SEC model for Parkinson’s disease. Science 295, 865–868.
Infected cortical cells (8,000,000 cells/10 cm plate) were harvested in Triton
Brady, R.O., Kanfer, J., and Shapiro, D. (1965). The metabolism of glucocere-
X-100 lysis buffer and 100,000 3 g Triton X-100 soluble lysate was loaded
brosides. I. Purification and properties of a glucocerebroside-cleaving enzyme
onto a Superdex 200 HR 10/300 column (GE healthcare, http://www.
from spleen tissue. J. Biol. Chem. 240, 39–43.
gelifesciences.com) as described previously (Mazzulli et al., 2006). Quantifica-
tion of a-syn oligomers has been described in detail previously (Tsika et al., 2010). Bultron, G., Kacena, K., Pearson, D., Boxer, M., Yang, R., Sathe, S., Pastores,
G., and Mistry, P.K. (2010). The risk of Parkinson’s disease in type 1 Gaucher
disease. J. Inherit. Metab. Dis. 33, 167–173.
a-Synuclein Protein Purification and Amyloid Measurements
Chandra, S., Gallardo, G., Fernandez-Chacon, R., Schluter, O.M., and Sudhof,
Recombinant human a-syn was purified from BL21 CodonPlus (DE3)-RIL
T.C. (2005). Alpha-synuclein cooperates with CSPalpha in preventing neuro-
competent E. coli (Agilent) as described previously (Mazzulli et al., 2007). Puri-
degeneration. Cell 123, 383–396.
fied a-syn was mixed with lipid dispersions (see Extended Experimental
Procedures) and amyloid formation was determined by thioflavin T binding Conway, K.A., Harper, J.D., and Lansbury, P.T. (1998). Accelerated in vitro
as described in the Supplemental Information. fibril formation by a mutant alpha-synuclein linked to early-onset Parkinson
disease. Nat. Med. 4, 1318–1320.
Histological Analysis of Gaucher Disease Mouse Models Cooper, A.A., Gitler, A.D., Cashikar, A., Haynes, C.M., Hill, K.J., Bhullar, B., Liu,
The homozygous point-mutated gba1 mice expressing V394L (4L) crossed to K., Xu, K., Strathearn, K.E., Liu, F., et al. (2006). Alpha-synuclein blocks ER-
the hypomorphic prosaposin mutant mice (PS-NA) have been described previ- Golgi traffic and Rab1 rescues neuron loss in Parkinson’s models. Science
ously (Sun et al., 2005). Histological analyses are described in the Extended 313, 324–328.
Experimental Procedures. Duda, J.E., Giasson, B.I., Mabon, M.E., Lee, V.M., and Trojanowski, J.Q.
(2002). Novel antibodies to synuclein show abundant striatal pathology in
Subcellular Fractionation Lewy body diseases. Ann. Neurol. 52, 205–210.
Infected cortical cells (8,000,000 cells/10 cm plate) were harvested in 0.25 M Erickson, A.H., Ginns, E.I., and Barranger, J.A. (1985). Biosynthesis of the lyso-
sucrose buffer containing 10 mM HEPES (pH 7.4) and 0.1M EDTA (SHB), somal enzyme glucocerebrosidase. J. Biol. Chem. 260, 14319–14324.
homogenized, and centrifuged at 6,800 3 g, 4 C, for 5 min. The remaining Giasson, B.I., Murray, I.V., Trojanowski, J.Q., and Lee, V.M. (2001). A hydro-
pellet was saved (P1). The supernatant was centrifuged at 17,000 3 g, 4 C, phobic stretch of 12 amino acid residues in the middle of alpha-synuclein is
for 10 min, supernatant removed (S), and the remaining pellet (P2) enriched essential for filament assembly. J. Biol. Chem. 276, 2380–2386.
in lysosomes was saved. Fraction S was centrifuged at 100,000 3 g for 1 hr
Giasson, B.I., Duda, J.E., Quinn, S.M., Zhang, B., Trojanowski, J.Q., and Lee,
to obtain P3. Pellets were extracted in 1% Triton X-100 lysis buffer, then 2%
V.M. (2002). Neuronal alpha-synucleinopathy with severe movement disorder
SDS buffer as described above. Fractions were analyzed by western blot anal-
in mice expressing A53T human alpha-synuclein. Neuron 34, 521–533.
ysis or by measuring GCase activity (Extended Experimental Procedures).
Goker-Alpan, O., Schiffmann, R., LaMarca, M.E., Nussbaum, R.L., McInerney-
Leo, A., and Sidransky, E. (2004). Parkinsonism among Gaucher disease
Statistical Analysis carriers. J. Med. Genet. 41, 937–940.
One-way ANOVA with Tukey’s post-hoc test was used in proteolysis, neuro-
Goker-Alpan, O., Giasson, B.I., Eblan, M.J., Nguyen, J., Hurtig, H.I., Lee, V.M.,
toxicity, immunostaining quantifications of LC3 and a-syn, P2 and P3 GCase
Trojanowski, J.Q., and Sidransky, E. (2006). Glucocerebrosidase mutations
activity assays, ANS, and thioflavin T determinations. One-way ANOVA with
are an important risk factor for Lewy body disorders. Neurology 67, 908–910.
Dunnet’s post-hoc test was used for post-ER/ER GCase ratios of cortical
neurons. Two-tailed Student’s t test was utilized for biochemical analyses, Grabowski, G.A. (2008). Phenotype, diagnosis, and treatment of Gaucher’s
quantification of a-syn and GCase protein levels, BODIPY 493 fluorescence disease. Lancet 372, 1263–1271.
analysis, and lipidomic analysis. p values less than 0.05 were considered Hodara, R., Norris, E.H., Giasson, B.I., Mishizen-Eberz, A.J., Lynch, D.R., Lee,
significant. Statistical calculations were performed with GraphPad Prism Soft- V.M., and Ischiropoulos, H. (2004). Functional consequences of alpha-synu-
ware, Version 4.0 (http://www.graphpad.com). clein tyrosine nitration: diminished binding to lipid vesicles and increased fibril
formation. J. Biol. Chem. 279, 47746–47753.
SUPPLEMENTAL INFORMATION Kayed, R., Head, E., Thompson, J.L., McIntire, T.M., Milton, S.C., Cotman,
C.W., and Glabe, C.G. (2003). Common structure of soluble amyloid oligomers
Supplemental Information includes Extended Discussion, Extended Experi- implies common mechanism of pathogenesis. Science 300, 486–489.
mental Procedures, seven figures, and six tables and can be found with this Lee, R.E. (1968). The fine structure of the cerebroside occurring in Gaucher’s
article online at doi:10.1016/j.cell.2011.06.001. disease. Proc. Natl. Acad. Sci. USA 61, 484–489.

Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc. 51

Lill, C.M., Roehr, J.T., McQueen, M.B., Bagade, S., Kavvoura, F., Schjeide, Sun, Y., Quinn, B., Witte, D.P., and Grabowski, G.A. (2005). Gaucher disease
B.M.M., Allen, N.C., Tanzi, R., Khoury, M.J., Ioannidis, J.P.A., and Bertram, mouse models: point mutations at the acid beta-glucosidase locus combined
L. (2008). The PDGene Database. Alzheimer Research Forum. Available at: with low-level prosaposin expression lead to disease variants. J. Lipid Res. 46,
http://www.pdgene.org Accessed 09/2008-03/2009. 2102–2113.
Manning-Bog, A.B., Schule, B., and Langston, J.W. (2009). Alpha-synuclein- Tayebi, N., Callahan, M., Madike, V., Stubblefield, B.K., Orvisky, E., Krasne-
glucocerebrosidase interactions in pharmacological Gaucher models: a bio- wich, D., Fillano, J.J., and Sidransky, E. (2001). Gaucher disease and parkin-
logical link between Gaucher disease and parkinsonism. Neurotoxicology sonism: a phenotypic and genotypic characterization. Mol. Genet. Metab.
30, 1127–1132. 73, 313–321.
Maroteaux, L., Campanelli, J.T., and Scheller, R.H. (1988). Synuclein: Tayebi, N., Walker, J., Stubblefield, B., Orvisky, E., LaMarca, M.E., Wong, K.,
a neuron-specific protein localized to the nucleus and presynaptic nerve Rosenbaum, H., Schiffmann, R., Bembi, B., and Sidransky, E. (2003). Gaucher
terminal. J. Neurosci. 8, 2804–2815. disease with parkinsonian manifestations: does glucocerebrosidase defi-
ciency contribute to a vulnerability to parkinsonism? Mol. Genet. Metab. 79,
Martinez, Z., Zhu, M., Han, S., and Fink, A.L. (2007). GM1 specifically interacts
104–109.
with alpha-synuclein and inhibits fibrillation. Biochemistry 46, 1868–1877.
Thayanidhi, N., Helm, J.R., Nycz, D.C., Bentley, M., Liang, Y., and Hay, J.C.
Mazzulli, J.R., Mishizen, A.J., Giasson, B.I., Lynch, D.R., Thomas, S.A., Naka-
(2010). Alpha-synuclein delays endoplasmic reticulum (ER)-to-Golgi transport
shima, A., Nagatsu, T., Ota, A., and Ischiropoulos, H. (2006). Cytosolic cate-
in mammalian cells by antagonizing ER/Golgi SNAREs. Mol. Biol. Cell 21,
chols inhibit alpha-synuclein aggregation and facilitate the formation of intra-
1850–1863.
cellular soluble oligomeric intermediates. J. Neurosci. 26, 10068–10078.
Trojanowski, J.Q., and Lee, V.M. (2002). Parkinson’s disease and related syn-
Mazzulli, J.R., Armakola, M., Dumoulin, M., Parastatidis, I., and Ischiropoulos,
ucleinopathies are a new class of nervous system amyloidoses. Neurotoxicol-
H. (2007). Cellular oligomerization of alpha-synuclein is determined by the
ogy 23, 457–460.
interaction of oxidized catechols with a C-terminal sequence. J. Biol. Chem.
282, 31621–31630. Tsika, E., Moysidou, M., Guo, J., Cushman, M., Gannon, P., Sandaltzopoulos,
R., Giasson, B.I., Krainc, D., Ischiropoulos, H., and Mazzulli, J.R. (2010).
Neudorfer, O., Giladi, N., Elstein, D., Abrahamov, A., Turezkite, T., Aghai, E., Distinct region-specific alpha-synuclein oligomers in A53T transgenic mice:
Reches, A., Bembi, B., and Zimran, A. (1996). Occurrence of Parkinson’s implications for neurodegeneration. J. Neurosci. 30, 3409–3418.
syndrome in type I Gaucher disease. QJM 89, 691–694.
Volles, M.J., and Lansbury, P.T., Jr. (2003). Zeroing in on the pathogenic form
Neumann, J., Bras, J., Deas, E., O’Sullivan, S.S., Parkkinen, L., Lachmann, of alpha-synuclein and its mechanism of neurotoxicity in Parkinson’s disease.
R.H., Li, A., Holton, J., Guerreiro, R., Paudel, R., et al. (2009). Glucocerebrosi- Biochemistry 42, 7871–7878.
dase mutations in clinical and pathologically proven Parkinson’s disease.
Wong, K., Sidransky, E., Verma, A., Mixon, T., Sandberg, G.D., Wakefield, L.K.,
Brain 132, 1783–1794.
Morrison, A., Lwin, A., Colegial, C., Allman, J.M., et al. (2004). Neuropathology
Seibler, P., Graziotto, J., Jeong, H., Simunovic, F., Klein, C., and Krainc, D. provides clues to the pathophysiology of Gaucher disease. Mol. Genet. Metab.
(2011). Mitochondrial parkin recruitment is impaired in neurons derived from 82, 192–207.
mutant PINK1 induced pluripotent stem cells. J. Neurosci. 31, 5970–5976.
Xu, Y.H., Quinn, B., Witte, D., and Grabowski, G.A. (2003). Viable mouse
Sidransky, E. (2005). Gaucher disease and parkinsonism. Mol. Genet. Metab. models of acid beta-glucosidase deficiency: the defect in Gaucher disease.
84, 302–304. Am. J. Pathol. 163, 2093–2101.
Sidransky, E., Nalls, M.A., Aasly, J.O., Aharon-Peretz, J., Annesi, G., Barbosa, Xu, Y.H., Sun, Y., Ran, H., Quinn, B., Witte, D., and Grabowski, G.A. (2010).
E.R., Bar-Shira, A., Berg, D., Bras, J., Brice, A., et al. (2009). Multicenter anal- Accumulation and distribution of alpha-synuclein and ubiquitin in the CNS of
ysis of glucocerebrosidase mutations in Parkinson’s disease. N. Engl. J. Med. Gaucher disease mouse models. Mol. Genet. Metab. 102, 436–437.
361, 1651–1661. Zala, D., Benchoua, A., Brouillet, E., Perrin, V., Gaillard, M.C., Zurn, A.D., Ae-
Stryer, L. (1965). The interaction of a naphthalene dye with apomyoglobin and bischer, P., and Deglon, N. (2005). Progressive and selective striatal degener-
apohemoglobin. A fluorescent probe of non-polar binding sites. J. Mol. Biol. ation in primary neuronal cultures using lentiviral vector coding for a mutant
13, 482–495. huntingtin fragment. Neurobiol. Dis. 20, 785–798.

52 Cell 146, 37–52, July 8, 2011 ª2011 Elsevier Inc.