a-Synuclein is part of a diverse and highly conserved
interaction network that includes PARK9 and
Aaron D Gitler1,2,6, Alessandra Chesi2,6, Melissa L Geddie1,6, Katherine E Strathearn3, Shusei Hamamichi4,
Kathryn J Hill5, Kim A Caldwell4, Guy A Caldwell4, Antony A Cooper5, Jean-Christophe Rochet3&
Parkinson’s disease (PD), dementia with Lewy bodies and multiple system atrophy, collectively referred to as synucleinopathies,
are associated with a diverse group of genetic and environmental susceptibilities. The best studied of these is PD. a-Synuclein
(a-syn) has a key role in the pathogenesis of both familial and sporadic PD, but evidence linking it to other predisposition factors
is limited. Here we report a strong genetic interaction between a-syn and the yeast ortholog of the PD-linked gene ATP13A2
(also known as PARK9). Dopaminergic neuron loss caused by a-syn overexpression in animal and neuronal PD models is rescued
by coexpression of PARK9. Further, knockdown of the ATP13A2 ortholog in Caenorhabditis elegans enhances a-syn misfolding.
These data provide a direct functional connection between a-syn and another PD susceptibility locus. Manganese exposure is an
environmental risk factor linked to PD and PD-like syndromes. We discovered that yeast PARK9 helps to protect cells from
manganese toxicity, revealing a connection between PD genetics (a-syn and PARK9) and an environmental risk factor (PARK9
and manganese). Finally, we show that additional genes from our yeast screen, with diverse functions, are potent modifiers of
a-syn–induced neuron loss in animals, establishing a diverse, highly conserved interaction network for a-syn.
Compelling evidence implicates a-syn in the pathogenesis of PD1,
including the identification of point mutations and locus duplica-
tion2,3and triplication4in familial forms, the abundance of a-syn in
Lewy bodies5and neurodegeneration resulting from increased expres-
sion of a-syn in multiple animal models6–9. Likewise, expression of
a-syn in yeast cells results in dosage-dependent toxicity10. Several
features of this toxicity, including production of reactive oxygen
species (ROS), lipid droplet accumulation and vesicle trafficking
defects, are reminiscent of a-syn toxicity in mammalian neurons10.
Therefore, yeast cells afford the opportunity to rapidly screen for
modifier genes with the hope that the identified genes will point to
common cellular mechanisms of toxicity and suggest avenues for
We recently reported the identification of a set of conserved genes
functioning in vesicular trafficking between the endoplasmic reticu-
lum (ER) and Golgi that are potent modifiers of a-syn toxicity in
yeast. Rab1, the mammalian ortholog of one of the encoded proteins
from the yeast screen, Ypt1, was tested in neuronal models of PD and
was able to prevent dopaminergic neuron loss12. We then expanded
this screen to include 5,000 yeast genes. In addition to vesicular
trafficking genes, we identified several other categories of a-syn–
toxicity modifiers, many with clear human orthologs, which reveal
additional complexities to synuclein pathology and which comple-
ment numerous studies of human synucleinopathies13.
A key question in the field is whether the genetic loci linked to PD
interact with each other or whether multiple independent insults
simply happen to result in a common phenotype (dopaminergic
neuron loss and resulting parkinsonism). There is emerging evidence
for a genetic interaction between PD-linked genes parkin and pink1 in
Drosophila14–16as well as interactions between a-syn and DJ-1,
another PD-linked protein17–20. Recently, the molecular nature of
the gene responsible for early-onset parkinsonism with pyramidal
degeneration and dementia (Kufor-Rakeb syndrome; MIM606693)
was elucidated and found to encode ATP13A2, a predicted lysosomal
P-type transmembrane cation transporting ATPase21–23.
As part of our unbiased screen to find modifiers of a-syn toxi-
city12,13, here we show that the yeast homolog of human ATP13A2
(PARK9) can suppress a-syn toxicity in yeast. This genetic interaction
between ATP13A2 and a-syn is conserved in neurons because
ATP13A2 expression in animal models of PD is sufficient to rescue
© 2009 Nature America, Inc. All rights reserved.
Received 1 October 2008; accepted 14 November 2008; published online 1 February 2009; doi:10.1038/ng.300
1Whitehead Institute for Biomedical Research and Howard Hughes Medical Institute, Cambridge, Massachusetts 02142, USA.2Department of Cell and Developmental
Biology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104, USA.3Department of Medicinal Chemistry and Molecular Pharmacology,
Purdue University, West Lafayette, Indiana 47907, USA.4Department of Biological Sciences, The University of Alabama, Tuscaloosa, Alabama 35487, USA.
5Garvan Institute of Medical Research, Sydney, NSW 2010, Australia.6These authors contributed equally to this work. Correspondence should be addressed to
NATURE GENETICS ADVANCE ONLINE PUBLICATION1
neurodegeneration. We also show that the yeast ortholog of the
ATP13A2 gene can protect cells from manganese toxicity, suggesting
an intimate connection between genetic and environmental causes of
neurodegeneration. Finally, we find other diverse modifiers of a-syn
toxicity from our screen, which are able to suppress a-syn–induced
neurodegeneration in animal and neuronal models of PD. These
include yeast genes and their human orthologs, encoding an E3
ubiquitin ligase (HRD1/SYVN1), ubiquitin protease (UBP3/USP10),
phosphodiesterase (PDE2/PDE9A), polo-like kinase (CDC5/PLK2)
and a casein kinase (YCK3/CSNK1G3). Thus, our data establish
a-syn as part of a highly conserved, multifaceted pathway.
Yeast ATP13A2 homolog suppresses a–syn toxicity
As part of our unbiased genetic screen for modifiers of a-syn
toxicity12, we discovered the yeast ortholog of human ATP13A2
(PARK9), an uncharacterized yeast gene designated YOR291W, to
be a suppressor of a-syn toxicity (Fig. 1a). We have therefore
named this yeast gene YPK9 (for Yeast PARK9). This and other
emerging work13–17,19suggest the possibility of many more connec-
tions between a-syn and other known causes of PD.
Knockout of YPK9 did not enhance a-syn toxicity (Fig. 1b).
Reasoning that this might be due to redundancy in function, we
carried out a BLAST search of the yeast genome and identified the
SPF1 (COD1) gene as encoding a highly related P-type ATPase (30%
identical, 49% similar, e value ¼ 3 ? 10?80). There are several P-type
ATPases in yeast, but SPF1 and YPK9 are by far the most closely
related to PARK9 (Supplementary Fig. 1 online). Overexpression of
SPF1 did not suppress a-syn toxicity (data not shown) and deletion of
SPF1 itself exerted a slight growth defect in yeast (Fig. 1b). However,
the SPF1 deletion was nearly lethal in cells expressing a single copy of
a-syn, which is by itself below the dosage threshold for toxicity in our
yeast model (Fig. 1b); the combination of the deletion of SPF1 and the
presence of a single copy of a-syn resulted in a growth defect that was
far greater than the sum of their individual toxicities. The double
knockout of SPF1 and YPK9 did not show a synthetic lethal phenotype
(Supplementary Fig. 2 online); however, overexpression of YPK9 was
sufficient to rescue the increase in a-syn toxicity caused by deletion of
SPF1 (data not shown), clearly establishing a functional overlap
between them. There are many possible explanations for the different
effects of deletion and overexpression of these genes. For example,
Spf1 expression or function could be feedback-regulated, preventing
overexpression from being efficacious, but leaving cells still vulnerable
to deletion. In any case, in yeast, the two genes most closely related
to human ATP13A2 have diverged such that YPK9 suppresses
a-syn toxicity when overexpressed whereas SPF1 enhances toxicity
Next, we investigated the relationship between Ypk9 and another
strong suppressor of a-syn toxicity from our screen, Ypt1 (ref. 12).
Ypt1 is the yeast homolog of human RAB1A, a guanosine tripho-
sphatase (GTPase) that regulates the trafficking of vesicles between the
endoplasmic reticulum (ER) and the Golgi. As previously reported,
overexpression of this protein rescues a-syn toxicity in both yeast and
neuronal cells12. To determine whether YPK9 and YPT1 suppress
a-syn toxicity in a mechanistically similar manner, we used a yeast
strain expressing higher levels of a-syn24. This HiTox strain shows
correspondingly higher toxicity and allows the detection of synergistic
effects between different suppressors that would not be possible
in the less toxic (IntTox) screening strain because YPT1 fully rescues
in that strain12,24.
In the HiTox strain, neither Ypt1 nor Ypk9 could suppress the
toxicity of a-syn. However, when Ypt1 and Ypk9 were both over-
expressed, we observed strong, albeit not complete, suppression
(Fig. 1c). Notably, this was not simply because the two proteins
provided additional levels of a redundant function that had a strong
threshold effect: cotransforming cells with two copies of either gene
itself did not rescue toxicity at all (Fig. 1c). Thus, the genetic
interaction between Ypt1 and Ypk9 is synergistic, indicating that
Ypk9 and Ypt1 function in mechanistically distinct ways to suppress
the toxic effects of a-syn accumulation.
In yeast and humans, the toxicity of a-syn is very dosage sensitive.
One way that Ypk9 might affect a-syn toxicity, therefore, would be to
© 2009 Nature America, Inc. All rights reserved.
CPY retained in ER (%)
Figure 1 Interaction between a-syn and the yeast
PARK9 homolog. (a) Spotting assays with yeast
a-syn toxicity modifier genes YPT1 and YPK9
showing their ability to suppress toxicity
compared to empty vector control. Fivefold serial
dilutions of yeast cells were spotted onto glucose
(a-syn expression repressed) or galactose (a-syn
expression induced). (b) Deletion of YPK9 has
no effect on a-syn toxicity; however, deleting the
closely related ATPase SPF1 enhances a-syn
toxicity. (c) Synergistic genetic interaction
between a-syn toxicity modifiers Ypt1 and Ypk9.
In a high toxicity (HiTox) two-copy a-syn yeast
strain, expression of Ypt1 or Ypk9 alone is not
sufficient to rescue toxicity. However, their
coexpression restores growth to this strain.
(d) Ypk9 overexpression eliminates a-syn
inclusions. Cells expressing a-Syn-YFP contain
many vesicular inclusions when transformed with
an empty vector and these are greatly diminished
in cells transformed with a Ypk9 expression
plasmid. (e) The ability of Ypk9 to suppress the
a-syn–induced block in ER-Golgi was measured
by carboxypeptidase Y (CPY) maturation assay.
Ypk9 considerably improved the trafficking of CPY
from ER to Golgi. Values represent means ± s.d.
2 ADVANCE ONLINE PUBLICATION NATURE GENETICS
alter its accumulation. However, both immunoblotting and fluores-
cent quantification of the a-syn fusion protein established that over-
expression of Ypk9 did not affect steady state levels of a-syn
(Supplementary Fig. 3 online). Ypk9 did, however, markedly alter
the localization of a-syn, largely restoring plasma membrane localiza-
tion and reducing intracellular inclusions (Fig. 1d). Previously,
inclusions formed by a-syn in yeast are associated with clusters of
mislocalized transport vesicles from various steps of the endocytic and
exocytic pathways24,25. Thus, these inclusions are a readout of vesicle
trafficking defects elicited by a-syn accumulation and may well relate
to early events in a-syn pathology seen in PD24,25. Our data suggest
that rescuing the vesicle trafficking block by overexpressing Ypt1 or
Ypk9 results in a reduction in the number of intracellular inclusions.
We next asked whether the two proteins would have similar effects
on the most immediate toxic defect that we have detected in cells
expressing a-syn, a defect in ER-to-Golgi trafficking12. We followed
carboxypeptidase Y (CPY) as it was trafficked through this pathway by
conducting a pulse-chase experiment. The subcellular location of CPY
is easily determined by compartment-specific glycosylations and
proteolytic cleavages that alter the molecular mass of the protein in
a well-characterized manner. a-Syn inhibits ER–Golgi transport and
prevents CPY from exiting the ER12(Fig. 1e). Ypk9 overexpression
significantly rescued the ability of proteins to leave the ER and traffic
to the Golgi (less protein in the ER, Fig. 1e), although the effect was
not as strong as that of Ypt1. Thus, Ypk9 and Ypt1 have mechan-
istically distinct functions, but both converge on vesicular transport to
antagonize a-syn toxicity.
that the intracellular
ATP13A2 homolog suppresses a-syn toxicity in C. elegans
To investigate the genetic relationship between a-syn and PARK9 in
dopaminergic neurons, a cell type directly relevant to human PD, we
turned first to the nematode model C. elegans (Fig. 2). Development
in the nematode is highly stereotyped and wild-type animals invari-
ably have exactly the same number of dopaminergic neurons. Expres-
sion of a-syn from the dopamine transporter (dat-1) gene promoter
resulted in an age-dependent progressive loss of dopaminergic neu-
rons7, with approximately 85% of animals having reduced numbers of
dopaminergic neurons at the 7-d stage (Fig. 2a,c). Expression of
W08D2.5 (catp-6, the C. elegans ATP13A2 ortholog) alone did not
induce any change in the number of dopaminergic neurons (data not
shown). Coexpression of W08D2.5 and a-syn, from the same pro-
moter (dat-1), partially rescued this neurodegeneration in each of four
independent transgenic lines (Fig. 2b,c).
We also used C. elegans to explore the consequences of ATP13A2
loss of function. Unfortunately, neuronal cells of this organism are
refractory to RNAi-mediated inhibition of gene expression26. How-
ever, our work with yeast and neuronal model systems establishes that
a-syn toxicity is the result of general cellular defects, to which
neuronal cells are simply more sensitive24. We therefore took advan-
tage of another cell type that has been extensively exploited for studies
of protein homeostasis in this organism27–30and is readily affected
As previously described, body-wall muscle cells that express a
human a-syn::GFP fusion show age-dependent a-syn aggregation
(Fig. 2d). Coexpression of tor-2, which encodes a chaperone protein
that reduces a-syn aggregation, provides a sensitized genetic back-
ground within which an enhancement of a-syn misfolding is readily
visualized27,31,32(Fig. 2e). In this sensitized background, we knocked
down the expression of the C. elegans ATP13A2 ortholog (W08D2.5)
by RNAi. This profoundly enhanced the misfolding of human a-syn
and did so in an age-dependent manner (Fig. 2f). Notably, it did so
without modifying the expression levels of a-syn or tor-2 (Supple-
mentary Fig. 4 online). These data provide further evidence for an
intimate functional interaction between PARK9 and a-syn.
ATP13A2 suppresses a-syn toxicity in primary neuron cultures
To validate our findings in mammalian dopaminergic neurons, we
used primary neuronal cultures prepared from the midbrain region of
rat embryos at stage 17. Although this assay is far more laborious than
those using stable tissue culture cell lines, the toxicity of a-syn in this
setting is more robust and reproducible, likely because the cells retain
apoptotic mechanisms that are lost in immortalized cell lines. Also,
unlike the nematode model, these cultures provide an opportunity to
assess toxicity to dopaminergic neurons relative to other neurons.
Transduction of these cells with lentivirus encoding a PD-linked
mutant a-syn (A53T) causes a reduction in the total number of
neurons (MAP2-positive staining), including those using g-amino
butyric acid (GABA-positive) as a neurotransmitter. But tyrosine
hydroxylase–positive dopaminergic neurons were even more severely
affected12,17,33. Co-transduction with a lentivirus encoding ATP13A2
(human PARK9) was potently protective. This was apparent from the
increased percentage of tyrosine hydroxylase–positive cells (Fig. 3a),
from the marked restoration of neuronal processes in cells expressing
tyrosine hydroxylase and from the restoration of more normal
neuronal morphology throughout the culture (Fig. 3b). ATP13A2
also increased the ratio of tyrosine hydroxylase–positive neurons
compared to MAP2-positive neurons (Fig. 3a). Thus, the relationship
© 2009 Nature America, Inc. All rights reserved.
Figure 2 PARK9 antagonizes a-syn-mediated
dopaminergic neuron degeneration in C. elegans.
(a,b) Anterior dopaminergic neurons in worms
expressing Pdat?1::GFP + Pdat?1::a-syn at the
day 7 stage. Arrowheads and arrows depict cell
bodies and neuronal processes, respectively.
Wild-type worms have six anterior dopaminergic
neurons. (a) a-Syn toxicity is depicted by the
loss of anterior dopaminergic neurons.
(b) Dopaminergic neurons are protected when
Pdat?1::FLAG-W08D2.5 cDNA is coexpressed.
(c) Quantification of C. elegans PARK9 rescue
of a-syn–induced neurodegeneration in four
independent transgenic lines displaying all six anterior dopaminergic neurons. *P o 0.05, Student’s t test. Values represent means ± s.d. (d) Overexpression
of a-syn in Punc?54::a-syn::GFP results in misfolding and aggregation of a-syn in body wall muscle cells at the young adult stage. (e) Co-overexpression of
TOR-2, a protein with chaperone activity, attenuates the misfolding of the a-syn::GFP protein. (f) The misfolding of a-syn::GFP is enhanced following RNAi
α-Syn + W08D2.5
Tg 2Tg 3
Worms with wild-type
dopaminergic neurons (%)
α-Syn + W08D2.5
α-Syn::GFP + TOR-2
+ W08D2.5 RNAi
α-Syn::GFP + TOR-2
NATURE GENETICS ADVANCE ONLINE PUBLICATION3
between PARK9 function and a-syn pathobiology that we had
discovered in yeast is conserved in mammals.
Ypk9 subcellular localization and effect of ATP13A2 mutations
Homozygous mutations in ATP13A2 have been identified as causing a
hereditary form of parkinsonism with dementia23. That both alleles
must be mutant to cause disease suggests that a recessive loss of
function is the root cause. However, ATP13A2 is expressed at a tenfold
higher level in the surviving neurons of the substantia nigra of subjects
with sporadic forms of PD23. Therefore, it was also reasonable to
suppose that the high expression of ATP13A2 in sporadic PD and the
two mutant alleles in familial forms represent a proteotoxic gain of
function with, in the latter case, two alleles required to cross a disease-
threshold burden23. The fact that in yeast overexpression of YPK9
suppressed a-syn toxicity supports the simpler view that it is a deficit
of PARK9 function that leads to disease. To explore this further,
we first determined the localization of the wild-type yeast and
We used homologous recombination to chromosomally tag Ypk9
with the yellow fluorescent protein (YFP). YFP-Ypk9, expressed from
its native promoter, localized strongly to the vacuole membrane
(Fig. 4a), consistent with the localization of the human protein to
the lysosome, the mammalian organellar equivalent of the yeast
vacuole23. We obtained similar results expressing GFP-Ypk9 fusion
proteins from a low-copy (CEN) plasmid with a constitutive promoter
(GPD) (Fig. 4a). Co-staining with a lipophilic dye, FM4-64, which
concentrates at the vacuole membrane, confirmed that this localiza-
tion was vacuolar (data not shown).
The human protein ATP13A2 also localized to the vacuole mem-
brane in yeast cells (Fig. 4a). However, even with a high-copy plasmid,
it was expressed at lower levels than the yeast protein. This is common
for multipass transmembrane proteins expressed across such large
evolutionary distances (ATP13A2 has ten predicted transmembrane
domains). Not unexpectedly, the human protein was unable to protect
against a-syn toxicity. Therefore, to test the effect of the human
mutations on yeast PARK9 localization and function, we took advan-
tage of the homology between the proteins to introduce equivalent
mutations into the yeast ortholog. Both forms of yeast YPK9 with
mutations implicated in human familial PD (‘subject-based’ muta-
tions) encoded proteins that were aberrantly localized. Ypk9 (D833-
1472) was expressed at lower levels than wild-type Ypk9 and was
distributed throughout the cytosol, in a punctate pattern, whereas
Ypk9 (D1329-1472) was retained in the ER (Fig. 4a).
Next we tested the ability of the YPK9 mutants to rescue a-syn
toxicity in our yeast model. Overexpression of wild-type Ypk9
suppressed toxicity, but the two altered Ypk9 proteins did not
(Fig. 4b). Moreover, expression of the altered Ypk9 proteins in
wild-type yeast cells (without a-syn expression) did not affect growth
(Fig. 4b), further supporting the notion that these are loss-of-function
and not dominant negative mutations. Human ATP13A2 and yeast
Ypk9 are predicted P-type ATPases34(Supplementary Fig. 3). We also
altered a conserved residue in Ypk9, predicted to abolish ATPase
© 2009 Nature America, Inc. All rights reserved.
Figure 3 PARK9 antagonizes a-syn–mediated
dopaminergic neuron degeneration in rat primary
midbrain cultures. (a) Human PARK9 (ATP13A2)
protects rat midbrain primary dopaminergic
neurons from a-synA53T–induced toxicity.
Primary rat embryonic midbrain cultures were
either mock infected (control) or infected with
lentivirus encoding LacZ, ATP13A2 alone,
a-synA53T alone or a-synA53T and ATP13A2.
Selective loss of dopaminergic neurons was
assessed immunocytochemically by determining
the percentage of MAP2-positive neurons that
also stained positive for tyrosine hydroxylase (TH).
N Z 3, #P o 0.05, ##P o 0.01, ###P o 0.001, one-way analysis of variance with Newman-Keuls post-test (a-synA53T versus control is also
P o 0.001). Values represent means ± s.d. (b) ATP13A2 rescues a-synA53T–induced dopaminergic neuron loss in rat primary midbrain cultures.
Representative micrographs of cells stained for MAP2 (red) and tyrosine hydroxylase (green). Arrows indicate dopaminergic neurons positive for both MAP2
and tyrosine hydroxylase. Scale bar, 20 mm.
TH-positive neurons (%)
α-SynA53T + ATP13A2
Glucose (α-syn ‘off’)Galactose (α-syn ‘on’)
Figure 4 Ypk9 is localized to the vacuole in yeast
and ATP13A2 subject-based mutations affect its
ability to rescue a-syn toxicity. (a) Fluorescence
microscopy to visualize Ypk9 subcellular
localization. A chromosomally tagged YFP fusion
(Ypk9-YFP) localizes to the vacuolar membrane,
as does wild-type GFP-Ypk9 expressed from the
constitutive GPD promoter. ATP13A2 mutations23
alter Ypk9 localization, but the ATPase-dead
mutant (D781N) does not. GFP-tagged human
ATP13A2 also localizes to the vacuole in yeast
cells. (b) Spotting assays with wild-type or a-syn–
expressing cells. Wild-type YPK9 overexpression
suppresses a-syn toxicity, but the two ATP13A2,
subject-based mutant YPK9 genes as well as the
ATPase-dead mutant do not. Expressing mutant
YPK9 in wild-type cells does not inhibit growth,
supporting the idea that these are loss-of-function
and not dominant-negative mutations.
4 ADVANCE ONLINE PUBLICATION NATURE GENETICS
activity (D781N). This resulting protein localized properly to the
vacuole (Fig. 4a) but was unable to rescue a-syn toxicity (Fig. 4b).
Taken together, our data indicate that both vacuolar localization and
ATPase activity are required for Ypk9 to antagonize a-syn toxicity.
ypk9D cells are hypersensitive to manganese
Little is known about the normal function of PARK9 or how it might
contribute to PD. To gain mechanistic insight into PARK9 function,
we explored the function of the yeast homolog. Both the yeast and
human proteins are predicted to be transmembrane cationic metal
transporters, but their substrate specificity has remained elusive23,34,35.
We tested several metals to identify potential functions for Ypk9. We
grew wild type and ypk9D cells in media containing a wide range of
metals and metal chelators at various concentrations to determine the
concentration that partially inhibited growth in our genetic back-
ground. This provided sensitized conditions to test the effects of Ypk9
(Fig. 5 and data not shown).
Of all the conditions we tested, ypk9D cells
were more sensitive to manganese (Mn2+)
than were wild-type cells (Fig. 5a,b,e). This
was detectable in rich media (Fig. 5a), and
even more so in minimal media (Fig. 5b),
and occurred in cells grown either on plates
or in liquid (Fig. 5e). Notably, ypk9D cells
were also slightly resistant to copper (Fig. 5a).
Expression of wild-type Ypk9 from an extra-
chromosomal plasmid with a strong promo-
ter was sufficient to rescue the Mn2+
sensitivity and indeed to make both ypk9D
and wild-type cells more resistant to Mn2+
(Fig. 5c). We were unable to complement this
phenotype with the human gene, which, as
noted above, we attribute to our inability to
express the human protein at sufficient levels
in our yeast system. We therefore used YPK9
with subject-based mutations to test the
effects of the human mutations on Ypk9’s
ability to protect against Mn2+toxicity.
Whereas wild-type Ypk9 suppressed Mn2+
toxicity, expression of the disease-associated
Ypk9 proteins did not (Fig. 5c). ATPase
activity was also required to protect against
Mn2+toxicity because the ATPase-dead Ypk9D781Nprotein failed to
rescue the defect (Fig. 5c). Moreover, the GFP-tagged Ypk9 fusion
proteins, which we used for the localization studies (Fig. 4a), were
functional, because they were able to rescue Mn2+sensitivity (Fig. 5d).
Thus, yeast Ypk9, and possibly human ATP13A2, likely function as
manganese transporters to protect cells from excess Mn2+exposure.
Validating additional genes from yeast screen in PD models
The other a-syn toxicity modifier genes we discovered in our yeast
screen13offer a multitude of promising possibilities for discovering
new therapeutic strategies. But it is axiomatic that this approach will
only work if hits from the yeast screen can be validated in neurons. As
an initial step toward this goal, we chose a subset of genes from our
screen for further analysis in neuronal PD models. Our sole criteria
were to test representative genes (i) from diverse functional categories,
(ii) with different strengths of suppression, (iii) with clear human
orthologs (Table 1) and (iv) with readily obtainable gene clones. We
© 2009 Nature America, Inc. All rights reserved.
0 5 10 15 20
WT (2 mM Mn2+)
ypk9∆ (2 mM Mn2+)
30 35 40
Figure 5 PARK9 protects cells from elevated
manganese levels. (a) Examples of conditions
used to identify the substrate specificity of YPK9.
We identified ypk9D cells as being sensitive to
manganese (Mn2+) relative to wild-type cells.
(b) The effect of various Mn2+concentrations
on ypk9D cells grown on rich (YPD) or synthetic
(CSM) media. (c) Expressing wild-type YPK9 in
ypk9D cells is sufficient to rescue Mn2+
sensitivity but neither the ATP13A2 subject-
based mutants nor the ATPase-dead mutant are
able to rescue. Expressing YPK9 in wild-type
yeast cells makes them more resistant to Mn2+
(compare top and bottom spottings). (d) YFP- (top
panel) and GFP-tagged (bottom panel) Ypk9
fusion proteins used for localization studies are
functional because they are able to protect
against Mn2+sensitivity. (e) ypk9D cells
also show sensitivity to Mn2+when grown in
NATURE GENETICS ADVANCE ONLINE PUBLICATION5
tested five suppressor genes, using human expression clones because
of their availability at the time the experiments were done (yeast/
human: HRD1/SYVN1, UBP3/USP10, PDE2/PDE9A, CDC5/PLK2,
YCK3/CSNK1G3), in the rat primary neuron lentiviral model and
the C. elegans a-syn model. Four out of five were efficacious. Two
(PLK2 and PDE9A) even suppressed a-syn–induced dopaminergic
neuron loss in the nematode (Supplementary Fig. 5 online). Because
we used human expression clones for these studies, it is not surprising
that the suppressors were more effective in rat neurons than in
nematode, which is separated from human by B800 million years.
In any case, these studies establish a highly conserved genetic inter-
action network operating between a-syn and several genes of diverse
function from yeast to mammals.
There are no clear homologs of a-syn in either yeast or nematode.
How relevant, then, is it to study this human PD gene in yeast? Very.
Work from our laboratory12,13,24and others25,36,37indicates that
a-syn, likely through its ability to bind lipids and associate with
membranes, is involved in the control of vesicle trafficking, a core
function conserved in all eukaryotes. We have discovered an a-syn
interaction network in yeast consisting of proteins with very diverse
functions (for example, kinases, phosphatases, metal transporters,
deubiquitinating enzymes)12,13and demonstrated its functional con-
servation in neurons of the rat and nematode, organisms separated
from yeast by a billion years of evolution. This notable and unexpected
degree of conservation not only confirms a conserved and fundamen-
tally important role for peripheral membrane proteins such as a-syn
in normal vesicle trafficking, but also indicates that this function is
deeply integrated with, and regulated by, other diverse and conserved
a-Syn is a very small (14 kDa) protein that binds lipids and is
peripherally associated with membranes. Although it folds when
associated with membranes38–40, it is otherwise natively unfolded,
with a propensity to form toxic oligomeric species41. We suggest that it
is these very basic properties of a-syn that account for the conserva-
tion of its pathobiology from yeast to man. Yeast cells may have
proteins with similar function. If so, they are likely constrained
more by protein–lipid than protein–protein interactions and have
simply diverged too greatly over the enormous evolutionary distances
covered here to allow clear recognition of functional homologs by
amino acid sequence.
The discoveries of single-gene mutations in familial forms of PD
over the last ten years provides an opportunity for further elucidating
the fundamental mechanisms of PD42. Our approach has revealed
genetic interactions between human disease genes, encoding a-syn
and PARK9, for which there was previously no known relationship.
The fact that five of the six genes that we discovered in yeast also affect
the toxicity of a-syn in neurons suggests that other modifiers
recovered in our screen will also be relevant13. It may, therefore, be
useful to test polymorphisms in these genes for association in
synucleinopathies. An additional challenge is to explore the complex-
ities of gene–gene and gene–environment interactions in PD. Our
identification of a connection between a-syn, PARK9 and manganese
also provides a toehold for these investigations.
A unifying theme emerging from our work, as well as that of several
other laboratories, is that a-syn sits at a nodal point, integrating a
lesions13,43,44. We hope the ability to dissect the nature of these
interactions, and how they contribute to disease in a variety of
model systems will forge new avenues for understanding disease
mechanisms and suggest therapeutic approaches aimed at the funda-
mental biological lesions in the complex disorders associated with
a-syn that intersect with diverse aspects of pathology.
Yeast strains and media. The a-synuclein–expressing yeast strainwe used in the
modifier screen was IntTox: a-syn-WT, MATa can1-100 his3-11,15 leu2-3,112
trp1-1 ura3-1 ade2-1 pRS303Gal-a-synWT-YFP pRS304Gal-aSynWT-YFP. The
a-synuclein expressing yeast strain we used for combinatorial gene analysis was
HiTox: a-syn-WT, MATa can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 ade2-1
reporter strain used to determine the effect of modifier genes on expression
from galactose-regulated promoter was Gal-YFP, MATa can1-100 his3-11,15
leu2-3,112 trp1-1 ura3-1 ade2-1 pRS303Gal-YFP. For Ypk9 localization studies,
a cassette containing a loxP-flanked KanMX cassette followed by YFP (pDH22,
a gift from Yeast Resource Center, University of Washington, Seattle) was
inserted in frame at the N terminus of Ypk9 by homologous recombina-
tion in the BY4741 strain background. Correct insertion was checked by
PCR and the KanMX gene was subsequently removed by transformation
with a plasmid containing a GAL-inducible Cre recombinase (pSH47, a gift
from Yeast Resource Center, University of Washington, Seattle). The ypk9D
strain was obtained by replacing the YPK9 coding region with the HIS3
gene in the BY4741 strain background. Colony PCR was used to verify
correct gene disruption. Strains were manipulated and media prepared using
The Gal promoter
Plasmids. For Ypk9 localization studies, pAG416GPD-EGFP-Ypk9 was con-
structed by Gateway cloning using the Ypk9 entry clone (pDONR221-YPK9)
and pAG416GPD-EGFP-ccdb destination vector45in an LR reaction. ATP13A2
subject-based mutations were introduced into pDONR221-YPK9 using the
QuikChange Site Directed Mutagenesis Kit (Stratagene) and sequence-verified.
Del833?1472corresponds to human mutation 1632_1653dup22 and Del1329?1472
corresponds to human mutation 3075delC. The ATPase-dead mutation
encodes D781N. Using Gateway cloning, we subcloned wild-type and mutant
YPK9 into pAG416GPD-EGFP-ccdB (for localization and Mn2+rescue),
pAG416GPD (Mn2+rescue) or pBY011 (a-synuclein rescue). Primer sequences
are available upon request. For studies with human PARK9, pcDNA3.1V5-
His-Topo-ATP13A2 was a gift from C. Kubisch (University of Cologne).
The ATP13A2 coding region was PCR-amplified and subcloned in pDONR221
to generate the entry clone pDONR221-ATP13A2. Subsequent LR Gateway
reactions generated pBY011-ATP13A2, pAG416GPD-ATP13A2 and pAG416
© 2009 Nature America, Inc. All rights reserved.
Table 1 a-Syn toxicity modifiers tested in neuronal PD models
Yeast GeneTypePredicted function Human orthologEffect in Yeast Effect in C. elegansEffect in rat neurons
E3 ubiquitin ligase
6 ADVANCE ONLINE PUBLICATION NATURE GENETICS
Metals. Serial dilutions of wild-type (BY4741) or ypk9D cells were spotted onto
YPD or CSM agar plates supplemented with excess concentrations of metals
(Ca2+, Fe3+, Mn2+, Zn2+, Co2+, Cu2+) or metal chelators (10 mM EGTA,
0.75 mM EDTA) and growth was assessed after 2–3 d at 30 1C. For the Mn2+
toxicity rescue experiments, wild-type and ypk9D strains were transformed with
the indicated plasmids and transformants spotted onto SD-URA plates con-
taining different concentrations of MnCl2(8, 10, 12, 14 mM). To assess ypk9D
Mn2+sensitivity in liquid culture, we used the Bioscreen to monitor growth.
Yeast cells were pre-grown in YPD to mid-log phase, diluted to OD600¼ 0.1
and dispensed to individual wells. OD600measurements were taken every 30
min in the presence of the indicated concentrations of MnCl2, and the plates
were shaken every 30 s to aerate the cells. At least three independent runs
were conducted for each growth condition, and each condition was tested
Phylogenetic tree. Protein sequences for all yeast and human P-type ATPases
were retrieved from the UniProtKB/Swiss-Prot family/domain classification
database (cation transport ATPase (P-type) family). A multiple sequence
alignment was obtained using the ClustalWalgorithm with default parameters.
The phylogenetic tree was obtained using the PROML program (maximum
likelihood algorithm with Jones-Taylor-Thornton probability model, constant
rate of change among sites) in the PHYLIP package (v3.67).
a-Syn toxicity modifier screen. We carried out the high-throughput yeast
transformation protocol as described previously for a smaller library
Combinatorial analysis. For the combinatorial analysis, pAG413GPD,
pAG413GPD-YPK9 or pAG413GPD-Ypt1 was cotransformed along with
pAG415GPD, pAG415GPD-YPK9 or pAG415GPD-Ypt1 into the HiTox
a-syn yeast strain using the standard lithium acetate technique. The transfor-
mants were plated onto SD-His/Leu agar plates and grown for 2 d. Cells were
then normalized and spotted onto SD-His/Leu and SGal-His/Leu plates.
Suppressors of a-syn–induced toxicity were identified on the galactose plates
after 3 d of growth at 30 1C.
ER-Golgi trafficking assay. We carried out the carboxypeptidase Y (CPY)
maturation assay as described previously12.
C. elegans experiments. Nematodes were maintained following the standard
procedures46. RNAi and fluorescent microscopy were done as described47by
feeding UA50 (baInl3; Punc?54::a-syn::gfp, Punc?54::tor-2, rol-6 (su1006)) worms
with the RNAi clones (Geneservice) corresponding to the worm orthologs of
YPK9 and its interactors. RNA isolation, cDNA preparation and semiquanti-
tative RT-PCR were conducted as described48with the following modification.
Total RNAs from 50 young-adult control (RNAi bacteria HT115(DE3) with
empty vector) and RNAi-treated worms were isolated to generate cDNAs. PCR
was then done using primers specific for amplifying cdk-5 as loading control,
a-syn and tor-2. For dopaminergic neurodegeneration analysis, strains UA51
(baEx42; Pdat?1::a-syn, Pdat?1::gfp, Pdat?1::FLAG-W08D2.5, rol-6 (su1006)) and
UA108 (baEx83; Pdat?1::gfp, Pdat?1::FLAG-W08D2.5, Punc?54::mCherry) were
generated by injecting 50 mg/ml of each expression plasmid into integrated
Pdat?1::a?syn, Pdat?1::GFP as well as Pdat?1::GFP worms, respectively. The
stable lines were analyzed for neurodegeneration as described previously7,12,24.
Rat primary midbrain neuron culture experiments. Primary midbrain
cultures were prepared, transduced with lentivirus and analyzed immunocyto-
chemically, as described previously12. All of the methods involving animal
handling were reviewed and approved by the Purdue Animal Care and Use
Committee. Relative dopaminergic cell viability was determined by counting
MAP2- and tyrosine hydroxylase–immunoreactive neurons in randomly cho-
sen observation fields. The data were expressed as the percentage of MAP2-
positive neurons that were also tyrosine hydroxylase–positive (this ratiometric
approach was used to correct for variations in cell density). Typically, 300–
1,500 MAP2-positive cells were counted per experiment for each condition. In
the control conditions, and in conditions where suppressors are efficacious, we
typically count 500–1,500 MAP2-positive neurons, a range that corresponds to
20–60 tyrosine hydroxylase–positive neurons. It is more difficult to obtain
such high cell counts from cultures expressing A53T a-synuclein alone
because the cell viability is markedly reduced. In these cases we typically
count 300–500 MAP2-positive neurons.
Preparation of primary mesencephalic cultures. Whole brains were dissected
from day 17 embryos obtained from pregnant Sprague-Dawley rats (Harlan).
The mesencephalic region containing the substantia nigra and ventral tegmental
area was isolated stereoscopically, and the cells were dissociated with trypsin
(final concentration, 26 mg/ml in 0.9% (w/v) NaCl). The cells were plated on
coverslips pretreated with poly-L-lysine (5 mg/ml) in media comprised of
DMEM, 10% (v/v) FBS, 10% (v/v) horse serum, penicillin (100 U/ml) and
streptomycin (100 mg/ml). After a 4-d incubation, the cells were treated for 48 h
with cytosine arabinoside (AraC) (20 mM) to suppress the growth of glial cells.
Methods involving animal handling were approved by the Purdue Animal Care
and Use Committee.
Preparation of lentiviral constructs. The ViraPower Lentivirus Expression
System (Invitrogen) was used to generate lentiviruses encoding human a-syn
(A53T), ATP13A2, CSNK1G3, USP10, PDE9A and PLK2 as described pre-
viously12. The insert from a pENTR-based entry construct was transferred into
the pLENTI6/V5 DEST lentiviral expression vector (Invitrogen) via recombina-
tion. The lentiviral construct was sequenced using an Applied Biosystems DNA
sequencer and packaged into virus via transient transfection of the 293FT
packaging cell line. We showed in a previous study that lentiviruses prepared
using this method have similar transduction efficiencies for MAP2- and tyrosine
hydroxylase–positive neurons (approximately 90% and 80%, respectively)17.
Note: Supplementary information is available on the Nature Genetics website.
We are grateful to C. Kubisch (University of Cologne) for providing the human
ATP13A2 cDNA and to the Yeast Resource Center for plasmids. A.D.G. was a
Lilly Fellow of the Life Sciences Research Foundation and is currently a Pew
Scholar in the Biomedical Sciences. A.D.G. is also supported by the US National
Institutes of Health Director’s New Innovator Award Program, part of the NIH
Roadmap for Medical Research, through grant number 1-DP2-OD004417-01.
A.C. is supported by a postdoctoral fellowship from the Parkinson’s Disease
Foundation. S.L. acknowledges support from the MGH/MIT Morris Udall Center
of Excellence in Parkinson Disease Research, NS038372, and the Howard Hughes
Medical Institute. M.L.G. was supported by a grant from the National Parkinson
Foundation. C. elegans studies in the Caldwell laboratory were supported in part
by grants from the Michael J. Fox Foundation, American Parkinson Disease
Foundation and Bachmann-Strauss Dystonia and Parkinson Foundation. Research
in the Rochet laboratory was supported by National Institutes of Health Grant
NS049221 and a grant from the American Parkinson Disease Association.
COMPETING INTERESTS STATEMENT
The authors declare competing financial interests: details accompany the full-text
HTML version of the paper at http://www.nature.com/naturegenetics/.
Published online at http://www.nature.com/naturegenetics/
Reprints and permissions information is available online at http://npg.nature.com/
1. Lee, V.M. & Trojanowski, J.Q. Mechanisms of Parkinson’s disease linked to pathological
alpha-synuclein: new targets for drug discovery. Neuron 52, 33–38 (2006).
2. Chartier-Harlin, M.C. et al. Alpha-synuclein locus duplication as a cause of familial
Parkinson’s disease. Lancet 364, 1167–1169 (2004).
3. Ibanez, P. et al. Causal relation between alpha-synuclein gene duplication and familial
Parkinson’s disease. Lancet 364, 1169–1171 (2004).
4. Singleton, A.B. et al. alpha-Synuclein locus triplication causes Parkinson’s disease.
Science 302, 841 (2003).
5. Spillantini, M.G. et al. Alpha-synuclein in Lewy bodies. Nature 388, 839–840
6. Auluck, P.K., Chan, H.Y., Trojanowski, J.Q., Lee, V.M. & Bonini, N.M. Chaperone
suppression of alpha-synuclein toxicity in a Drosophila model for Parkinson’s disease.
Science 295, 865–868 (2002).
7. Cao, S., Gelwix, C.C., Caldwell, K.A. & Caldwell, G.A. Torsin-mediated protection from
cellular stress in the dopaminergic neurons of Caenorhabditis elegans. J. Neurosci. 25,
8. Masliah, E. et al. Dopaminergic loss and inclusion body formation in alpha-synuclein
mice: implications for neurodegenerative disorders. Science 287, 1265–1269 (2000).
© 2009 Nature America, Inc. All rights reserved.
NATURE GENETICS ADVANCE ONLINE PUBLICATION7
9. Lo Bianco, C., Ridet, J.L., Schneider, B.L., Deglon, N. & Aebischer, P. alpha- Download full-text
Synucleinopathy and selective dopaminergic neuron loss in a rat lentiviral-based
model of Parkinson’s disease. Proc. Natl. Acad. Sci. USA 99, 10813–10818 (2002).
10. Outeiro, T.F. & Lindquist, S. Yeast cells provide insight into alpha-synuclein biology and
pathobiology. Science 302, 1772–1775 (2003).
11. Gitler, A.D. Beer and bread to brains and beyond: can yeast cells teach us about
neurodegenerative disease? Neurosignals 16, 52–62 (2008).
12. Cooper, A.A. et al. Alpha-synuclein blocks ER-Golgi traffic and Rab1 rescues neuron
loss in Parkinson’s models. Science 313, 324–328 (2006).
13. Yeger-Lotem, E. et al. Nat. Genet. (in the press).
14. Clark, I.E. et al. Drosophila pink1 is required for mitochondrial function and interacts
genetically with parkin. Nature 441, 1162–1166 (2006).
15. Park, J. et al. Mitochondrial dysfunction in Drosophila PINK1 mutants is complemen-
ted by parkin. Nature 441, 1157–1161 (2006).
16. Yang, Y. et al. Mitochondrial pathology and muscle and dopaminergic neuron degen-
eration caused by inactivation of Drosophila Pink1 is rescued by Parkin. Proc. Natl.
Acad. Sci. USA 103, 10793–10798 (2006).
17. Liu, F., Nguyen, J.L., Hulleman, J.D., Li, L. & Rochet, J.C. Mechanisms of DJ-1
neuroprotection in a cellular model of Parkinson’s disease. J. Neurochem. 105, 2435–
18. Meulener, M.C. et al. DJ-1 is present in a large molecular complex in human brain
tissue and interacts with alpha-synuclein. J. Neurochem. 93, 1524–1532 (2005).
19. Batelli, S. et al. DJ-1 modulates alpha-synuclein aggregation state in a cellular model
of oxidative stress: relevance for Parkinson’s disease and involvement of HSP70. PLoS
ONE 3, e1884 (2008).
20. Zhou, W. & Freed, C.R. DJ-1 up-regulates glutathione synthesis during oxidative stress
and inhibits A53T alpha-synuclein toxicity. J. Biol. Chem. 280, 43150–43158
21. Di Fonzo, A. et al. ATP13A2 missense mutations in juvenile parkinsonism and young
onset Parkinson disease. Neurology 68, 1557–1562 (2007).
22. Lees, A.J. & Singleton, A.B. Clinical heterogeneity of ATP13A2 linked disease (Kufor-
Rakeb) justifies a PARK designation. Neurology 68, 1553–1554 (2007).
23. Ramirez, A. et al. Hereditary parkinsonism with dementia is caused by mutations in
ATP13A2, encoding a lysosomal type 5 P-type ATPase. Nat. Genet. 38, 1184–1191
24. Gitler, A.D. et al. The Parkinson’s disease protein alpha-synuclein disrupts cellular Rab
homeostasis. Proc. Natl. Acad. Sci. USA 105, 145–150 (2008).
25. Soper, J.H. et al. a-Synuclein induced aggregation of cytoplasmic vesicles in Sacchar-
omyces cerevisiae. Mol. Biol. Cell 19, 1093–1103 (2008).
26. Kennedy, S., Wang, D. & Ruvkun, G. A conserved siRNA-degrading RNase negatively
regulates RNA interference in C. elegans. Nature 427, 645–649 (2004).
27. Caldwell, G.A. et al. Suppression of polyglutamine-induced protein aggregation in
Caenorhabditis elegans by torsin proteins. Hum. Mol. Genet. 12, 307–319
28. Cohen, E., Bieschke, J., Perciavalle, R.M., Kelly, J.W. & Dillin, A. Opposing activities
protect against age-onset proteotoxicity. Science 313, 1604–1610 (2006).
29. Link, C.D. Expression of human beta-amyloid peptide in transgenic Caenorhabditis
elegans. Proc. Natl. Acad. Sci. USA 92, 9368–9372 (1995).
30. Satyal, S.H. et al. Polyglutamine aggregates alter protein folding homeostasis in
Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 97, 5750–5755 (2000).
31. McLean, P.J. et al. TorsinA and heat shock proteins act as molecular chaperones:
suppression of alpha-synuclein aggregation. J. Neurochem. 83, 846–854 (2002).
32. Sharma, N. et al. A close association of torsinA and alpha-synuclein in Lewy bodies: a
fluorescence resonance energy transfer study. Am. J. Pathol. 159, 339–344 (2001).
33. Outeiro, T.F. et al. Sirtuin 2 inhibitors rescue a-Synuclein-mediated toxicity in models
of Parkinson’s disease. Science 317, 516–519 (2007).
34. Kuhlbrandt, W. Biology, structure and mechanism of P-type ATPases. Nat. Rev. Mol.
Cell Biol. 5, 282–295 (2004).
35. Axelsen, K.B. & Palmgren, M.G. Evolution of substrate specificities in the P-type
ATPase superfamily. J. Mol. Evol. 46, 84–101 (1998).
36. Gosavi, N., Lee, H.J., Lee, J.S., Patel, S. & Lee, S.J. Golgi fragmentation occurs in the
cells with prefibrillar alpha-synuclein aggregates and precedes the formation of fibrillar
inclusion. J. Biol. Chem. 277, 48984–48992 (2002).
37. Larsen, K.E. et al. Alpha-synuclein overexpression in PC12 and chromaffin cells
impairs catecholamine release by interfering with a late step in exocytosis. J. Neurosci.
26, 11915–11922 (2006).
38. Kubo, S. et al. A combinatorial code for the interaction of alpha-synuclein with
membranes. J. Biol. Chem. 280, 31664–31672 (2005).
39. Weinreb, P.H., Zhen, W., Poon, A.W., Conway, K.A. & Lansbury, P.T. Jr. NACP, a protein
implicated in Alzheimer’s disease and learning, is natively unfolded. Biochemistry 35,
40. Eliezer, D., Kutluay, E., Bussell, R. Jr. & Browne, G. Conformational properties of
alpha-synuclein in its free and lipid-associated states. J. Mol. Biol. 307, 1061–1073
41. Volles, M.J. & Lansbury, P.T. Jr. Zeroing in on the pathogenic form of alpha-synuclein
and its mechanism of neurotoxicity in Parkinson’s disease. Biochemistry 42,
42. Forman, M.S., Trojanowski, J.Q. & Lee, V.M. Neurodegenerative diseases: a decade of
discoveries paves the way for therapeutic breakthroughs. Nat. Med. 10, 1055–1063
43. Norris, E.H. et al. Pesticide exposure exacerbates alpha-synucleinopathy in an A53T
transgenic mouse model. Am. J. Pathol. 170, 658–666 (2007).
44. Dauer, W. et al. Resistance of alpha -synuclein null mice to the parkinsonian neurotoxin
MPTP. Proc. Natl. Acad. Sci. USA 99, 14524–14529 (2002).
45. Alberti, S., Gitler, A.D. & Lindquist, S. A suite of Gateway((R)) cloning vectors for
high-throughput genetic analysis in Saccharomyces cerevisiae. Yeast 24, 913–919
46. Brenner, S. The genetics of Caenorhabditis elegans. Genetics 77, 71–94 (1974).
47. Kamath, R.S. & Ahringer, J. Genome-wide RNAi screening in Caenorhabditis elegans.
Methods 30, 313–321 (2003).
48. Hamamichi, S. et al. Hypothesis-based RNAi screening identifies neuroprotective genes
in a Parkinson’s disease model. Proc. Natl. Acad. Sci. USA 105, 728–733 (2008).
© 2009 Nature America, Inc. All rights reserved.
8 ADVANCE ONLINE PUBLICATION NATURE GENETICS