LRRK2 Parkinson disease mutations enhance its microtubule association.

Abstract

Dominant missense mutations in leucine-rich repeat kinase 2 (LRRK2) are the most common genetic causes of Parkinson disease (PD) and genome-wide association studies identify LRRK2 sequence variants as risk factors for sporadic PD. Intact kinase function appears critical for the toxicity of LRRK2 PD mutants, yet our understanding of how LRRK2 causes neurodegeneration remains limited. We find that most LRRK2 PD mutants abnormally enhance LRRK2 oligomerization, causing it to form filamentous structures in transfections of cell lines or primary neuronal cultures. Strikingly, ultrastructural analyses, including immuno-electron microscopy and electron microscopic tomography, demonstrate that these filaments consist of LRRK2 recruited onto part of the cellular microtubule network in a well-ordered, periodic fashion. Like LRRK2-related neurodegeneration, microtubule association requires intact kinase function and the WD40 domain, potentially linking microtubule binding and neurodegeneration. Our observations identify a novel effect of LRRK2 PD mutations and highlight a potential role for microtubules in the pathogenesis of LRRK2-related neurodegeneration.

Full text

Page 1

LRRK2 Parkinson disease mutations enhance
its microtubule association
Lauren R. Kett1,3,{, Daniela Boassa4,5,{, Cherry Cheng-Ying Ho7,{, Hardy J. Rideout8,
Junru Hu4, Masako Terada4, Mark Ellisman4,5,6,∗and William T. Dauer1,2,∗
1Department of Neurology,2Department of Cell and Developmental Biology and3Department of Neuroscience
Graduate Program, University of Michigan Medical School, Ann Arbor, MI, USA,4National Center for Microscopy and
Imaging Research, Center for Research in Biological Systems and5Department of Neurosciences, University of
California, San Diego, CA, USA,6Affiliate Professor, Institute for Systems Biology, Seattle, WA, USA,7Department
of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA and8Biomedical Research
Foundation, Academy of Athens, Athens, Greece
Received July 25, 2011; Revised September 29, 2011; Accepted November 7, 2011
Dominant missense mutations in leucine-rich repeat kinase 2 (LRRK2) are the most common genetic causes
of Parkinson disease (PD) and genome-wide association studies identify LRRK2 sequence variants as risk
factors for sporadic PD. Intact kinase function appears critical for the toxicity of LRRK2 PD mutants, yet
our understanding of how LRRK2 causes neurodegeneration remains limited. We find that most LRRK2
PD mutants abnormally enhance LRRK2 oligomerization, causing it to form filamentous structures in trans-
fections of cell lines or primary neuronal cultures. Strikingly, ultrastructural analyses, including immuno-
electron microscopy and electron microscopic tomography, demonstrate that these filaments consist of
LRRK2 recruited onto part of the cellular microtubule network in a well-ordered, periodic fashion. Like
LRRK2-related neurodegeneration, microtubule association requires intact kinase function and the WD40
domain, potentially linking microtubule binding and neurodegeneration. Our observations identify a novel
effect of LRRK2 PD mutations and highlight a potential role for microtubules in the pathogenesis of
LRRK2-related neurodegeneration.
INTRODUCTION
The currently understood biology of leucine-rich repeat kinase
2 (LRRK2) suggests that studies of this protein may provide
new insights into neurodegeneration that are broadly relevant
tosporadic Parkinson disease(PD)and amenableto therapeutic
targeting. Genome-wide association studies demonstrate that
common variation around the locus that encodes LRRK2 segre-
gates with increased risk for PD (1,2), and missense mutations
in LRRK2 cause a clinical and neuropathological syndrome
that is indistinguishable from typical-appearing sporadic PD
(3). LRRK2 contains GTPase and kinase domains, as well
as leucine-rich repeat (LRR) and WD40 protein–protein
interaction domains (Fig. 1A). Many potentially pathogenic
sequence alterations have been identified in LRRK2, but only
five missense mutations (Fig. 1A) clearly segregate with PD
in large family studies (4). Two of these mutations (R1441G,
R1441C) are located in the GTPase domain (termed Ras of
complex proteins, or ‘Roc’ domain), a third (Y1699C) falls in
a region between the GTPase and kinase domains (termed the
C-terminal of Roc, or ‘COR’ domain) and two other mutations
(G2019S and I2020T) are in the kinase domain.
LRRK2 appears to exist as a dimer (5–7), and studies of
fragments of LRRK2 or its prokaryotic homolog indicate
that dimerization occurs in the Roc-COR region (5). Structural
analyses of these fragments indicate that the R1441C PD
mutation (or mutation of the analogous residue in the prokary-
otic protein) can destabilize the dimer formed by these frag-
ments (5,8). Yet, available data do not define the functional
significance of LRRK2 self-association and it is unclear
†Contributed equally to this publication.
∗To whom correspondence should be addressed. Tel: +1 8585342251; Email: mark@ncmir.ucsd.edu (M.E.); Tel:+1 7346153874; Email: dauer@
umich.edu (W.T.D.)
# The Author 2011. Published by Oxford University Press. All rights reserved.
For Permissions, please email: journals.permissions@oup.com
Human Molecular Genetics, 2012, Vol. 21, No. 4
doi:10.1093/hmg/ddr526
Advance Access published on November 11, 2011
890–899
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 2

Figure 1. Multiple pathogenic mutations enhance LRRK2 oligomerization and filament formation in a kinase-dependent manner. (A) Domain structure and
Parkinson’s disease mutations of LRRK2. LRR, leucine-rich repeat; GTP, GTPase domain (also called Roc domain, Ras of complex proteins). Five PD-causing
missense mutations are shown. (B–E) The formation of LRRK2 filaments is enhanced by multiple PD mutations. Cath.a-differentiated (CAD) cells were trans-
fected with WT (B) or PD mutant forms (D; R1441C, R1441G, Y1699C, G2019S, I2020T) of GFP-tagged LRRK2. WT-LRRK2 adopted either a diffuse, aggre-
gated or filamentous pattern of subcellular localization (B; labeled using anti-GFP antibody). The frequency of cells bearing LRRK2 aggregates or filaments was
quantified 48 h after transfection (C and E). Multiple PD mutations increased the percentage of cells with LRRK2 filaments (labeled using anti-GFP antibody).
Data are means+SE of four to five independent experiments (∗∗P , 0.01,∗∗∗P , 0.001, n.s., non-significant; ANOVA with Tukey’s test). (F) Expression of
LRRK2 in neuronal processes. Yellow fluorescent protein (YFP)-LRRK2-Y1699C was transfected into primary cortical neurons and imaged by confocal mi-
croscopy (left image, YFP fluorescence shown in green). A similar pattern of filament formation was observed in neurites in primary neurons transfected
with untagged LRRK2-I2020T (right image, labeled using an anti-LRRK2 antibody shown in red). (G) WT LRRK2 oligomerizes and multiple LRRK2 PD muta-
tions enhance its oligomerization. V5-LRRK2 was co-expressed in CAD cells with GFP or GFP-LRRK2. Lysates were immunoprecipitated with anti-GFP 48 h
after transfection, and the immunoprecipitates were analyzed with anti-V5 and anti-GFP immunoblots. The oligomeric state of LRRK2 is shown as a relative
ratio of co-purified V5-LRRK2 to GFP-LRRK2, normalized to the WT-LRRK2 ratio.
Human Molecular Genetics, 2012, Vol. 21, No. 4 891
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 3

whether this property is altered in PD mutant forms of the full-
length protein. One possibility is that LRRK2 self-association
regulates its kinase activity, but only the G2019S mutation
clearly increases kinase activity (by ?3–5-fold), whereas
the other mutations appear to have little or no effect on
kinase function (9–13), at least in the in vitro assays used
thus far. Intact kinase function does appear necessary for
LRRK2 toxicity in vitro and in vivo (9,12,14). LRRK2-
induced neurodegeneration of primary neuronal cultures
is caspase-dependent, and may involve activation of the fas-
associated protein with death domain (FADD)–caspase-8
pathway (15). One site of LRRK2 toxicity may be in neuronal
processes, as overexpression of LRRK2 in vitro or in vivo
causes neurite shortening, whereas the loss of LRRK2 function
leads to increased neuron length and branching (11). The mo-
lecular details that underlie LRRK2 effects in neuronal pro-
cesses are not clear, but may involve interactions with the
cytoskeleton or cytoskeletal-related molecules (10,16–18).
Here, we provide evidence that four PD mutations (R1441C,
R1441G, Y1699C, I2020T) cause LRRK2 to decorate microtu-
bules in a well-ordered periodic fashion, as evidenced by
immunofluorescence, immuno-electron microscopy (EM) and
EM tomographic studies. Microtubule-associated LRRK2
appears as filamentous structures in transfected cell lines or
primary neuronalcultures, and the frequency ofthese structures
increases with microtubule stabilization (taxol) and decreases
with microtubule dissolution
LRRK2-inducedneurodegeneration,
requires intact kinase function and the WD40 domain, poten-
tially linking LRRK2-microtubule association to neurodegen-
eration. Our observations identify a novel effect of LRRK2
PD mutations, and provide a platform to further dissect
LRRK2-related signals by identifying factors that modulate
the interaction between LRRK2 and microtubules.
(nocodazole).Similar
formation
to
filament
RESULTS
LRRK2 PD mutations enhance LRRK2 filament formation
In multiple cell lines (including Cath.a-differentiated (CAD),
human embryonic kidney (HEK) and HeLa) transfected with
wild-type (WT) LRRK2, we observed that LRRK2 adopted a
primarily diffuse cytosolic distribution, but we also observed
LRRK2-positive aggregates and a distinct pattern of filament
formation (Fig. 1B). These patterns were observed with differ-
ent epitope tags (e.g. GFP and V5) as well as with untagged
protein (data not shown). To determine whether PD mutations
alter the subcellular distribution of LRRK2, we quantified the
frequency of these patterns (diffuse, aggregate or filament)
for WT and all PD mutant alleles (Fig. 1C and E). This analysis
wasperformedintheneuronalcatecholaminergicCADcellline
because of the robust transfection efficiency possible with this
system. The percentage of cells with punctate aggregates was
not increased by any of the PD mutant alleles (Fig. 1C), so
these structures may represent a non-specific effect of
LRRK2 overexpression. In contrast, filament formation was
significantly enhanced by all LRRK2 PD mutations that do
not consistently enhance kinase function (9,10,13), whereas
no effect on filament formation was observed for the single
mutant, G2019S, that consistently and significantly increases
kinase activity (Fig. 1D and E; Supplementary Material,
Fig. S1; 9,10,13). Notably, steady-state protein levels of WT
and PD mutant LRRK2 are similar (data not shown), so these
filament-forming mutations do not appear to act by altering
LRRK2 stability. Morphologically, these filaments are reminis-
cent of death-effector filaments formed by FADD, caspase-8,
tumor necrosis factor receptor type 1-associated death domain
protein and BCL10 (19–21). This observation intrigued us
because all of these filament-forming proteins participate in
cell death-related signaling, potentially providing a clue to
LRRK2 function relevant to neurodegeneration. We found
that LRRK2 filaments also occur in LRRK2-transfected
primary neurons, both in the cell body and in neurites, and
with both yellow fluorescent protein (YFP)-tagged as well as
untagged protein (Fig. 1F).
For previously characterized filament-forming proteins,
filament formation reflects a homotypic protein–protein
interaction that is required for their normal signaling function
(19,20,22). Oligomerization is also a signaling mechanism
employed by RIP1 and RIP2 kinases (23), close phylogenetic
relatives of LRRK2. These facts led us to hypothesize that the
filaments represent enhanced LRRK2 oligomerization. To
investigate whether LRRK2 oligomerizes, we co-expressed
GFP- and V5-LRRK2 in CAD cells and tested whether they
co-immunoprecipitate (co-IP). Lysates were immunoprecipi-
tated with a GFP antibody and probed for associated
V5-LRRK2. Differentially tagged WT-LRRK2 molecules did
co-purify, indicating that LRRK2 can oligomerize (Fig. 1G).
Moreover, LRRK2 oligomerization was clearly enhanced by
filament-forming PD mutations (R1441C, Y1699C, I2020T),
whereas the non-filament-inducing mutant (G2019S) did
not differ from the WT-LRRK2 control (Fig. 1G). The oligo-
merization status of LRRK2 mutants therefore correlates
withtheleveloffilamentformation,indicatingthatthecytosolic
filaments are related to this feature of LRRK2 biology.
Because the LRRK2 filaments appear similar to cytoskeletal
structures, we explored whether LRRK2 was templating onto
existing cytoskeletal elements or forming novel ‘LRRK2 only’
structures. To determine which, if any, cytoskeletal structures
LRRK2 may interact with, we examined cells transfected with
FLAG-LRRK2-I2020T with both anti-FLAG and a panel of
antibodies against the major cytoskeletal elements, as well
as with mito-tracker, since mitochondria exist in filamentous
strands. However, LRRK2 filament staining did not colocalize
with any of these markers (data not shown).
LRRK2 decorates microtubules in an organized manner
To further explore the nature of LRRK2 filaments, we
employed EM to define the ultrastructure of these structures.
We performed correlated light and EM by transfecting
HEK293T cells with YFP-LRRK2-I2020T, using YFP fluores-
cence to identify filament-bearing cells, and then processed the
same samples for EM to obtain high-resolution information on
LRRK2 distribution in specific subcellular domains. Figure 2
shows filament formation identified by YFP fluorescence
(Fig. 2A–C) and the corresponding low-magnification elec-
tron micrograph of a thin section from the same area
(Fig. 2D). Higher magnifications (Fig. 2E and F) reveal
filamentous structures that are organized in parallel arrays.
892Human Molecular Genetics, 2012, Vol. 21, No. 4
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 4

The filaments consist of microtubules, identified by their char-
acteristic structure in high-resolution EM as well as the corre-
lated immunofluorescence, and these consist of LRRK2
apparently recruited onto parts of these unusually arrayed
microtubules (Fig. 2F and G). We observed areas of naked
microtubules (indicated by black arrows in Fig. 2G) inter-
rupted by bundles of microtubules decorated by electron dens-
ities around them (indicated by black arrowheads in Fig. 2G).
Immuno-EM using a specific antibody to LRRK2 confirmed
the specificity of LRRK2 densities around the microtubules
(Fig. 3A). To further examine the 3D electron microscopic
Figure 3. Immuno-EM and electron tomography of LRRK2 filaments.
(A) Immuno-EM was performed with a LRRK2-specific antibody. Immuno-
gold signals appear as black dots (indicated by white arrows) on bundles of
microtubules. For comparison, white arrowheads indicate a microtubule with
the absence of LRRK2 staining. (B–E) Electron tomography of LRRK2 fila-
ments. Higher magnification images of non-decorated tubules (B) and deco-
rated microtubules (C) from electron tomograms are shown as well as a low
power field from the area reconstructed by electron tomography (D) from
which a graphical reconstruction was produced using Amira (E). These data
are available for viewing and download from the Cell Centered Database
either as images or 3D movies.
Figure 2. Pattern of expression of LRRK2 using correlated light and EM.
(A–C) The distribution of LRRK2 filaments was revealed in HEK293T
cells expressing YFP-LRRK2-I2020T by the YFP fluorescence, (D) correlated
image at low-power EM and (E–G) intermediate magnifications. Intermediate
magnification of filaments (G) shows areas of naked microtubules (arrows) as
well as electron densities (arrowheads) around them.
Human Molecular Genetics, 2012, Vol. 21, No. 4893
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 5

organization of the filaments, we performed electron tomog-
raphy on 250 nm thick sections. The tomograms show that
LRRK2 appears to interact closely with the microtubules in
a well-ordered, periodic fashion (Fig. 3B–E).
The tomographic data showing LRRK2 ‘coating’ microtu-
bules suggested to us that masking of the tubulin epitope by
LRRK2 may explain the lack of co-localization between
tubulin and LRRK2 immunofluorescence (noted above).
Indeed, a re-examination of individual Z-images from the
immunofluorescent staining (rather than the maximal intensity
projection) showed distinct regions in which LRRK2 filaments
appear as direct continuations of microtubules (Fig. 4A–G).
Moreover, we further explored the LRRK2-microtubule associ-
ation by testing whether the frequency of filament formation
was altered by microtubule stability. To accomplish this, we
treated LRRK2-transfected CAD cells with the microtubule-
stabilizing drug taxol (10 mM for 5 h), the microtubule-
destabilizing drug nocodazole (100 nM for 1 h) or vehicle
[dimethyl sulfoxide (DMSO)] control. An examiner blinded
to genotype and drug treatment then counted the proportion
of cells with filaments. We found a significant effect of
genotype, drug treatment and the interaction between the
two [two-way analysis of variance (ANOVA): F ¼ 14.36,
P , 0.0001]. In cells transfected with either GFP-LRRK2-
I2020T or Y1699C, taxol significantly increased the percentage
of filament-bearing cells, whereas this percentage was signifi-
cantly decreased by nocodozole (Fig. 4H). In contrast, taxol
or nocodazole did not significantly change the percentage of
filament-bearing cells transfected with GFP-LRRK2-WT or
G2019S. This distinction between the behavior of the
G2019S and the other PD mutations is similar to that observed
for their kinase activity (Supplementary Material, Fig. S1; 9–
13). Taken together with the observations of LRRK2-coating
microtubules by EM tomography, these data further indicate
that multiple LRRK2 PD mutations (but not G2019S)
enhance its ability to associate with microtubules. As cellular
stressors, it is possible that taxol and nocodazole treatment
alter LRRK2 behavior through indirect mechanisms, although
thefactthattheymodulatefilamentformationinoppositedirec-
tions makes this unlikely.
LRRK2-microtubule association requires intact kinase
function and the WD40 domain
Because intact kinase activity is required for the enhanced
neurotoxicity of LRRK2 PD mutants (9,12; Supplementary
Material, Fig. S1), we tested whether this activity was also
required for the enhanced microtubule association caused by
LRRK2 PD mutations. Consistent with the neurotoxicity
data, blocking kinase activity by introducing a second muta-
tion in the kinase domain, K1906R, completely abolished
the effect of LRRK2 PD mutations on microtubule association
(Fig. 5A). To further elucidate the mechanism whereby
LRRK2 PD mutations provoke enhanced microtubule associ-
ation, we attempted to define a microtubule association
domain by performing a structure–function study, testing
which regions of LRRK2 are required (or are dispensable)
for filament formation. We used the I2020T mutant as a repre-
sentative mutation for these studies since the behavior of this
mutant does not differ from R1441C or Y1699C (Figs 1 and
4). This analysis demonstrated that the WD40 domain was
necessary for filament formation (Fig. 5B), whereas the
entire N-terminus (prior to the LRR domain) is dispensable
for microtubule binding. We reported previously (7) that
DWD40-LRRK2 is unable to autophosphorylate yet retains
the ability to trans-phosphorylate the model substrate MBP,
suggesting a potential relationship between autophosphoryla-
tion and microtubule association. Such a relationship could
contribute to an explanation of why blocking kinase function
virtually abolishes microtubule association (Fig. 5A).
DISCUSSION
Our study is the first to provide evidence that multiple PD
mutations enhance the association of LRRK2 with microtu-
bules. Several characteristics of this interaction indicate that
it may contribute to LRRK2-dependent neurodegeneration.
First, similar to LRRK2 neurotoxicity, this effect requires
intact LRRK2 kinase function and the presence of the
WD40 domain (7,9,12,14). Secondly, our colocalization and
ultrastructural analyses show that LRRK2 interacts closely
with microtubules in a well-ordered, periodic fashion, suggest-
ing the existence of a LRRK2-binding site on microtubules or
microtubule-bound proteins. Conversely, the specific require-
ment of the WD40 domain (but not the N-terminus) for
microtubule-bindingindicates
LRRK2 mediate microtubule association. Our studies also
show that LRRK2 filaments appear similar to linear arrays
of electron dense structures observed in human post-mortem
substantia nigra tissue from a patient carrying the filament-
forming Y1699C mutation (24). With the development of
better LRRK2 antibodies for immunohistochemical studies,
it will be interesting to see if similar filaments are observed
in the brains of PD patients with LRRK2 mutations.
The link between visible microtubule association (fila-
ments) and toxicity mirrors those observed for protein
aggregates seen with synuclein and other neurodegenerative-
related proteins: neurotoxicity occurs in the population of
LRRK2-transfected neurons, whereas visible filaments are
seen in a minority of cells. In all of these cases, the visible
protein lesions (filaments or aggregates) appear to reflect
toxic events that also occur in soluble (and therefore not
visualizable) protein. Our co-IP results (Fig. 1G) support
such a scenario since these studies, which assess the entire
cell population (including untransfected cells), show a
markedly increased association for filament-forming LRRK2
PD mutations.
Similarfilamentshave
Abarrategui et al. (25) for R1441C-LRRK2. That report
suggested that LRRK2 filaments were a form of multivesicu-
lar bodies, yet used a LRRK2 antibody for immuno-EM that
has not been validated using LRRK2 knockout tissue, and no
multivesicular body markers were shown to colocalize with
LRRK2 filaments. In contrast, we used a validated LRRK2
antibody for detection (as well as an epitope tag), and
employed correlated light and EM to ensure that we were
observing the ultrastructural features of the same structures
we visualized using immunofluorescence. We never observed
membrane in association with LRRK2 filaments, either by
that specificregions of
beenreportedbyAlegre-
894 Human Molecular Genetics, 2012, Vol. 21, No. 4
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 6

EM or using membrane dyes on LRRK2-transfected cells
(data not shown).
Dzamko et al. (26) reported that the treatment of LRRK2-
transfected HEK293 cells with the kinase inhibitor H-1152
(which inhibits the LRRK2 kinase) caused G2019S-LRRK2
to colocalize with microtubules, a finding that would appear
to be inconsistent with our finding that intact kinase function
is required for microtubule association. However, key differ-
ences preclude comparison of these studies. H-1152 is not a
LRRK2-specific inhibitor, and is therefore likely to inhibit a
variety of kinases in living cells. Thus, the localization
effect that Dzamko et al. report may be the indirect effect of
Figure 4. LRRK2 filaments associate with microtubules and are modified by microtubule-altering drugs. (A–G) LRRK2 filaments localize to microtubules.
HEK293TcellstransfectedwithFLAG-LRRK2-I2020Tweredoublelabeledwithanti-FLAG(DandG)andanti-alphatubulin(CandF)antibodies.LRRK2filaments
appearascontinuationsofmicrotubules(A,BandE,mergedimagesshowingtubulininred,LRRK2-FLAGingreenandnucleiinblue;insetsinBareshowninE,Fand
G). Arrows denote portions of microtubules positive for LRRK2 staining, but lacking tubulin staining. Scale bar is 10 mm. (H) Treatment with microtubule-altering
drugs changed the filament proportion. CAD cells were transfected with WT or PD mutant forms (G2019S, I2020T or Y1699C) of GFP-LRRK2. Forty-eight hours
post-transfection, cells were treated with the microtubule-stabilizing drug taxol (10 mM for 5 h), the microtubule-destabilizing drug nocodazole (100 nM for 1 h) or
dimethyl sulfoxide (DMSO) control. Following drug treatment, cells were fixed, stained for GFP and the frequency of cells with filaments was quantified. Data are
means+SEofthreeindependentexperiments.(∗P , 0.05,∗∗∗P , 0.001,n.s.,non-significant;two-wayanalysisofvariance(ANOVA)withpost-hocTukey’stest).
Human Molecular Genetics, 2012, Vol. 21, No. 4895
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 7

inhibiting a different kinase, and may not reflect inhibition of
the LRRK2 kinase per se. Moreover, Dzamko et al. do not
quantify the apparent effect of H-1152 treatment on subcellu-
lar localization, which is necessary to confirm suspected
effects, since a variable low rate of filament formation is
seen with both wild-type and G2019S-LRRK2 (Fig. 1E).
Notably, we do not observe an effect of H-1152 on the subcel-
lular localization of wild-type or PD mutant forms of LRRK2
(data not shown).
The interaction of LRRK2 with microtubules could affect
microtubule dynamics or microtubule-related transport, or
microtubules could serve as a scaffold to concentrate
LRRK2 signaling events such as the recruitment of the
FADD/caspase-8 complex. In vitro, LRRK2 can bind
alpha/beta-tubulin heterodimers (16) and can phosphorylate
beta-tubulin (17), and active (but not kinase dead) LRRK2
enhances the polymerization of tubulin (isolated from bovine
brain) in the presence of microtubule-associated proteins
(17). Further support for LRRK2 interaction with the
microtubule cytoskeleton includes the findings of deficient
microtubule associated protein Tau phosphorylation in brain
lysates from LRRK2 null mice (17), and multiple reports of
alterations of neurite length and branching in situations of
LRRK2 deficiency or overexpression (11,27–30). LRRK2 is
reported to signal via the FADD/caspase-8 pathway (15),
and in olfactory receptor neurons, degeneration of the axon
Figure 5. Kinase activity and WD40 domain are required for LRRK2-microtubule association. (A) The microtubule-association effect of LRRK2 PD mutations
is kinase-dependent. Blocking LRRK2 kinase function blocks the effect of filament formation caused by PD mutations. CAD cells were transfected with WT
LRRK2-GFP or ‘double mutant’ constructs containing a LRRK2 PD mutation and a kinase-deficient mutation (K1906R) and stained for filaments with an
anti-GFP antibody. Quantification of filament formation was done as in Figure 1E. Data are means+SE of three independent experiments (∗∗P , 0.01, n.s.:
non-significant; ANOVA with Tukey’s test). (B) Filament formation requires the WD40 domain of LRRK2. CAD cells were transfected with full-length
GFP-LRRK2-I2020T (FL-IT), GFP-LRRK2-I2020T lacking the WD40 domain (DWD40) or the Roc and WD40 domains of LRRK2-I2020T (Roc-WD-IT).
Forty-eight hours after transfection, the cells were assessed for filament formation. Anti-GFP immunoblot of transfected cell lysate demonstrates that truncated
proteins are produced at the same or greater level of expression as full-length LRRK2.
896 Human Molecular Genetics, 2012, Vol. 21, No. 4
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 8

and soma require synaptically activated caspase-8 that is retro-
gradely transported along microtubules (31). LRRK2 might
therefore modulate the axonal transport of caspase-8 or other
cargo, a notion consistent with the potentially central role of
axonal degeneration in PD (32).
One potential mechanism for the enhanced microtubule
association of mutant LRRK2 is that PD mutations may
directly enhance the affinity of LRRK2 for microtubules or
microtubule-bound proteins. Alternatively, PD mutations may
enhance LRRK2 homo-oligomerization or complex formation
with other cytosolic proteins, and these homo-oligomers or
protein complexes may themselves have enhanced affinity for
microtubules. Our data provide evidence that multiple patho-
genic mutations enhance LRRK2–LRRK2 interactions, an
effect that requires intact kinase activity and that is correlated
with filament formation. Studies from Wittinghofer and collea-
gues (8) suggest that the COR domain is important for LRRK2
dimerization, and data from Cookson and colleagues (5)
suggest that PD mutations that replace the arginine at position
1441 with either cysteine or glycine (which cause filaments)
weaken the interaction in the Roc-COR region. The structural
alteration produced by the weakening of this interaction
may expose a protein–protein interaction domain, providing
a potential mechanism for how PD mutations enhance
LRRK2-microtubule association. The fact that multiple muta-
tions provoke LRRK2 filaments indicates that other insults,
such as those that occur in idiopathic PD, might also promote
abnormal LRRK2 association with microtubules.
MATERIALS AND METHODS
Plasmids
LRRK2 cDNA from HEK 293 cells was cloned in pCMS-
EGFP (Clontech), pcDNA-DEST53 and -DEST40 (Invitro-
gen). All subsequent mutants were generated using site-
directed mutagenesis and all mutant clones were re-sequenced
to confirm their accuracy.
Cell lines and primary cortical neuron cultures
CAD cells were grown in Dulbecco’s modified eagle medium
(DMEM)/F12 (GIBCO) supplemented with 8% fetal bovine
serum. HEK293T cells were grown in DMEM (GIBCO)
with 10% serum. CAD cells were transfected with Lipofecta-
mine/PLUS (Invitrogen), whereas 293T cells were transfected
with Lipofectamine 2000 (Invitrogen) according to the manu-
facturer’s instructions. Cultures of cortical neurons from E16
mice were maintained in the Neurobasal medium containing
B-27 supplements (GIBCO), and transfected with Lipofecta-
mine 2000 four days after being plated. Primary neurons
were transfected with LRRK2 expression constructs and
pCMS-EGFP at 10:1 ratio. Each experiment was performed
on cover slips in triplicate three times, with .100 cells on
each cover slip being quantified.
Antibodies
Weusedthefollowingantibodies:rabbitanti-GFP(Abcam,Cat.
No. ab6656); mouse anti-GFP (Roche, Cat. No. 11814460001);
mouseanti-V5(Invitrogen,Cat.No.R960-25);rabbitmonoclo-
nalanti-LRRK2(Epitomics,Cat.No.3514-1);mousemonoclo-
nal anti-alpha tubulin (Sigma, Cat. No. T6074); and rabbit
polyclonal anti-FLAG (Abnova, Cat. No. PAB0900). The
rabbit anti-LRRK2 used to immunostain untagged LRRK2
was a kind gift from Dr Benoit Giasson (University of
Pennsylvania, Philadelphia, PA, USA).
Cell preparation for EM
HEK293T cells transfected with YFP-LRRK2-IT were fixed
for 30 min in 2% glutaraldehyde in 0.1 M sodium cacodylate
buffer (pH 7.4) on ice. After washes in 0.1 M cacodylate
buffer, cells were post-fixed with 1% osmium tetraoxide for
30 min, stained with 2% uranyl acetate and then dehydrated
and embedded in Durcupan epoxy resin. Sections were cut
using a diamond knife at a thickness of 70–90 nm for thin
sections and 250 nm for thick sections. Post-staining of grids
was done with 2% uranyl acetate for 5 min and lead citrate
for 1 min. Thin sections were examined using a JEOL 1200
FX1 operated at 120 kV.
Electron tomography
Colloidal gold particles (10 nm diameter) were deposited on
opposite sides of the thick epoxy sections to serve as fiducial
markers. For stability in the beam, the section was coated with
carbon on both sides. For reconstruction, a triple series of
images at regular tilt (angular increments of 28 from 2608
to +608 increments) was collected with a JEOL 4000EX
intermediate-voltageelectron
400 kV. The specimens were irradiated before initiating a
tilt series to limit anisotropic specimen thinning during
image collection. Tilt series were recorded using a slow-scan
CCD camera. The pixel dimensions of the CCD camera
were 4000 × 4000 and the pixel resolution was 0.754 nm.
Fine alignment and reconstruction was performed using the
TxBR software package (33).
microscope operatedat
Post-embedding immuno-EM
For ultrastructural analysis of the filaments, we used an
anti-LRRK2 antibody using 10 nm gold particles to identify
the LRRK2 protein around the microtubules. After labeling,
grids were washed and post-stained with 2% uranyl acetate
for 5 min and lead citrate for 1 min.
Immunofluorescent labeling
Forty-eight hours after transfection, methanol- or formalde-
hyde-fixed HEK293T or CAD cells on poly-D-lysine-coated
glass cover slips were blocked in phosphate buffered saline
(PBS) containing 0.3% Triton X-100, 0.5% bovine serum
albumin, 50 mM glycine and 2% normal donkey serum for
30 min. Cover slips were then incubated overnight at 48C in
primary antibodies diluted in block solution. The next day,
cover slips were washed, incubated with Alexa488- or
Alexa647-conjugated secondary antibodies and washed in
PBS before mounting using ProLong Gold antifade reagent
with4′,6-diamidino-2-phenylindole (Invitrogen).Data
Human Molecular Genetics, 2012, Vol. 21, No. 4897
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 9

acquisition was done with an Olympus FluoView1000
laser-scanning confocal microscope (Olympus, Center Valley,
PA, USA), using a 60X1.42 NA lens or with a ×100
oil-immersion objective with a Zeiss LSM510 2-photon con-
focal microscope. Image analysis of z-scan was done using
the Imaris Software (Bitplane AG, St Paul, MN, USA) and
the Image J software (rsb.info.nih.gov/ij).
Drug treatment
Forty-eight hours after transfection, cells on cover slips were
treated with taxol (10 mM in DMSO for 5 h), nocodazole
(100 nM in DMSO for 1 h) or DMSO. Cells were then
stained as described above.
co-IP analysis
CAD cells or 293T cells transfected with various expression
constructs were Dounce homogenized in lysis buffer (20 mM
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, pH 7.4,
150 mM NaCl, 0.1–0.5% NP-40, 2 mM ethylene glycol tetraa-
cetic acid, 2 mM MgCl2, 10% glycerol, 1 mM sodium orthova-
nadate, 10 mM NaF, 25 mM b-glycerophosphate, pH 7.2 and
protease inhibitors). After centrifugation at 20 000g for
15 min, the supernatants were pre-cleared with protein-A
agarose (Roche) for 30 min. Lysates containing 2 mg total
protein were immunoprecipitated with rabbit anti-GFP for
1 h–overnight followed by incubation with protein-A beads
for 2 h at 48C.
SUPPLEMENTARY MATERIAL
Supplementary Material is availabe at HMG online.
ACKNOWLEDGEMENTS
Thanks to Kana Tsukamoto, Dr Guido Gaietta and Ying Jones
for outstanding technical assistance. This work is also linked
to activities of a consortium of researchers working with the
Institute for Systems Biology in Seattle on Parkinson’s
disease.
Conflict of Interest statement. None declared.
FUNDING
This work was supported by National Institute of Neurological
Diseases and Stroke grant R01 NS061098-02 (W.D.), National
Institutes of Health P41 RR004050 (M.H.E.) and the Branf-
man Family Foundation (M.H.E.).
REFERENCES
1. Satake, W., Nakabayashi, Y., Mizuta, I., Hirota, Y., Ito, C., Kubo, M.,
Kawaguchi, T., Tsunoda, T., Watanabe, M., Takeda, A. et al. (2009)
Genome-wide association study identifies common variants at four loci as
genetic risk factors for Parkinson’s disease. Nat. Genet., 41, 1303–1307.
2. Simon-Sanchez, J., Schulte, C., Bras, J.M., Sharma, M., Gibbs, J.R., Berg,
D., Paisan-Ruiz, C., Lichtner, P., Scholz, S.W., Hernandez, D.G. et al.
(2009) Genome-wide association study reveals genetic risk underlying
Parkinson’s disease. Nat. Genet., 41, 1308–1312.
3. Zimprich, A., Biskup, S., Leitner, P., Lichtner, P., Farrer, M., Lincoln, S.,
Kachergus, J., Hulihan, M., Uitti, R.J., Caine, D.B. et al. (2004) Mutations
in LRRK2 cause autosomal-dominant parkinsonism with pleomorphic
pathology. Neuron, 44, 601–607.
4. Bonifati, V. (2007) LRRK2 low-penetrance mutations (Gly2019Ser) and
risk alleles (Gly2385Arg)-linking familial and sporadic Parkinson’s
disease. Neurochem. Res., 32, 1700–1708.
5. Deng, J., Lewis, P.A., Greggio, E., Sluch, E., Beilina, A. and Cookson,
M.R. (2008) Structure of the ROC domain from the Parkinson’s
disease-associated leucine-rich repeat kinase 2 reveals a dimeric GTPase.
Proc. Natl Acad. Sci. USA, 105, 1499–1504.
6. Greggio, E., Zambrano, I., Kaganovich, A., Beilina, A., Taymans, J.M.,
Daniels, V., Lewis, P., Jain, S., Ding, J., Syed, A. et al. (2008) The
Parkinson disease-associated leucine-rich repeat kinase 2 (LRRK2) is a
dimer that undergoes intramolecular autophosphorylation. J. Biol. Chem.,
283, 16906–16914.
7. Jorgensen, N.D., Peng, Y., Ho, C.C., Rideout, H.J., Petrey, D., Liu, P. and
Dauer, W.T. (2009) The WD40 domain is required for LRRK2
neurotoxicity. PLoS ONE, 4, e8463.
8. Gotthardt, K., Weyand, M., Kortholt, A., Van Haastert, P.J. and
Wittinghofer, A. (2008) Structure of the Roc-COR domain tandem of
C. tepidum, a prokaryotic homologue of the human LRRK2 Parkinson
kinase. EMBO J., 27, 2352.
9. Greggio, E., Jain, S., Kingsbury, A., Bandopadhyay, R., Lewis, P.,
Kaganovich, A., van der Brug, M.P., Beilina, A., Blankinton, J., Thomas,
K.J. et al. (2006) Kinase activity is required for the toxic effects of mutant
LRRK2/dardarin. Neurobiol. Dis., 23, 329–341.
10. Jaleel, M., Nichols, R.J., Deak, M., Campbell, D.G., Gillardon, F., Knebel,
A. and Alessi, D.R. (2007) LRRK2 phosphorylates moesin at
threonine-558: characterization of how Parkinson’s disease mutants affect
kinase activity. Biochem. J., 405, 307–317.
11. MacLeod, D., Dowman, J., Hammond, R., Leete, T., Inoue, K. and
Abeliovich, A. (2006) The familial Parkinsonism gene LRRK2 regulates
neurite process morphology. Neuron, 52, 587–593.
12. Smith, W.W., Pei, Z., Jiang, H., Dawson, V.L., Dawson, T.M. and Ross,
C.A. (2006) Kinase activity of mutant LRRK2 mediates neuronal toxicity.
Nat. Neurosci., 9, 1231–1233.
13. West, A.B., Moore, D.J., Choi, C., Andrabi, S.A., Li, X., Dikeman, D.,
Biskup, S., Zhang, Z., Lim, K.L., Dawson, V.L. et al. (2007) Parkinson’s
disease-associated mutations in LRRK2 link enhanced GTP-binding and
kinase activities to neuronal toxicity. Hum. Mol. Genet., 16, 223–232.
14. Lee, B.D., Shin, J.H., VanKampen, J., Petrucelli, L., West, A.B., Ko, H.S.,
Lee, Y.I., Maguire-Zeiss, K.A., Bowers, W.J., Federoff, H.J. et al. (2010)
Inhibitors of leucine-rich repeat kinase-2 protect against models of
Parkinson’s disease. Nat. Med., 16, 998–1000.
15. Ho, C.C., Rideout, H.J., Ribe, E., Troy, C.M. and Dauer, W.T. (2009) The
Parkinson disease protein leucine-rich repeat kinase 2 transduces death
signals via Fas-associated protein with death domain and caspase-8 in a
cellular model of neurodegeneration. J. Neurosci., 29, 1011–1016.
16. Gandhi, P.N., Wang, X., Zhu, X., Chen, S.G. and Wilson-Delfosse, A.L.
(2008) The Roc domain of leucine-rich repeat kinase 2 is sufficient for
interaction with microtubules. J. Neurosci. Res., 86, 1711–1720.
17. Gillardon, F. (2009) Leucine-rich repeat kinase 2 phosphorylates brain
tubulin-beta isoforms and modulates microtubule stability—a point of
convergence in parkinsonian neurodegeneration? J. Neurochem., 110,
1514–1522.
18. Parisiadou, L., Xie, C., Cho, H.J., Lin, X., Gu, X.L., Long, C.X.,
Lobbestael, E., Baekelandt, V., Taymans, J.M., Sun, L. et al. (2009)
Phosphorylation of ezrin/radixin/moesin proteins by LRRK2 promotes the
rearrangement of actin cytoskeleton in neuronal morphogenesis.
J. Neurosci., 29, 13971–13980.
19. Fotin-Mleczek, M., Henkler, F., Samel, D., Reichwein, M., Hausser, A.,
Parmryd, I., Scheurich, P., Schmid, J.A. and Wajant, H. (2002) Apoptotic
crosstalk of TNF receptors: TNF-R2-induces depletion of TRAF2 and IAP
proteins and accelerates TNF-R1-dependent activation of caspase-8.
J. Cell Sci., 115, 2757–2770.
20. Guiet, C. and Vito, P. (2000) Caspase recruitment domain
(CARD)-dependent cytoplasmic filaments mediate bcl10-induced
NF-kappaB activation. J. Cell Biol., 148, 1131–1140.
21. Siegel, R.M., Martin, D.A., Zheng, L., Ng, S.Y., Bertin, J., Cohen, J. and
Lenardo, M.J. (1998) Death-effector filaments: novel cytoplasmic
898Human Molecular Genetics, 2012, Vol. 21, No. 4
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from

Page 10

structures that recruit caspases and trigger apoptosis. J. Cell Biol., 141,
1243–1253.
22. Muppidi, J.R., Lobito, A.A., Ramaswamy, M., Yang, J.K., Wang, L., Wu,
H. and Siegel, R.M. (2006) Homotypic FADD interactions through a
conserved RXDLL motif are required for death receptor-induced
apoptosis. Cell Death Differ., 13, 1641–1650.
23. Inohara, N., Koseki, T., Lin, J., del Peso, L., Lucas, P.C., Chen, FF.,
Ogura, Y. and Nunez, G. (2000) An induced proximity model for
NF-kappa B activation in the Nod1/RICK and RIP signaling pathways.
J. Biol. Chem., 275, 27823–27831.
24. Wszolek, Z.K., Vieregge, P., Uitti, R.J., Gasser, T., Yasuhara, O.,
McGeer, P., Berry, K., Caine, D.B., Vingerhoets, F.J., Klein, C. et al.
(1997) German-Canadian family (Family A) with parkinsonism,
amyotrophy, and dementia—longitudinal observations. Parkinson. Relat.
Disord., 3, 125–139.
25. Alegre-Abarrategui, J., Christian, H., Lufino, M.M., Mutihac, R., Venda,
L.L., Ansorge, O. and Wade-Martins, R. (2009) LRRK2 regulates
autophagic activity and localizes to specific membrane microdomains in a
novel human genomic reporter cellular model. Hum. Mol. Genet., 18,
4022–4034.
26. Dzamko, N., Deak, M., Hentati, F., Reith, A.D., Prescott, A.R., Alessi,
D.R. and Nichols, R.J. (2010) Inhibition of LRRK2 kinase activity leads
to dephosphorylation of Ser(910)/Ser(935), disruption of 14–3–3
binding and altered cytoplasmic localization. Biochem. J., 430,
405–413.
27. Cookson, M.R. (2010) The role of leucine-rich repeat kinase 2 (LRRK2)
in Parkinson’s disease. Nat. Rev. Neurosci., 11, 791–797.
28. Dachsel, J.C., Behrouz, B., Yue, M., Beevers, J.E., Melrose, H.L. and
Farrer, M.J. (2010) A comparative study of Lrrk2 function in primary
neuronal cultures. Parkinson. Relat. Disord., 16, 650–655.
29. Lin, C.H., Tsai, P.I., Wu, R.M. and Chien, C.T. (2010) LRRK2 G2019S
mutation induces dendrite degeneration through mislocalization and
phosphorylation of tau by recruiting autoactivated GSK3b. J. Neurosci.,
30, 13138–13149.
30. Plowey, E.D., Cherra, S.J. 3rd, Liu, Y.J. and Chu, C.T. (2008) Role of
autophagy in G2019S-LRRK2-associated neurite shortening in
differentiated SH-SY5Y cells. J. Neurochem., 105, 1048–1056.
31. Carson, C., Saleh, M., Fung, F.W., Nicholson, D.W. and Roskams, A.J.
(2005) Axonal dynactin p150Glued transports caspase-8 to drive
retrograde olfactory receptor neuron apoptosis. J. Neurosci., 25,
6092–6104.
32. Cheng, H.C., Ulane, C.M. and Burke, R.E. (2010) Clinical progression
inParkinsondiseaseandtheneurobiologyofaxons.Ann.Neurol.,67,715–725.
33. Lawrence, A., Bouwer, J.C., Perkins, G. and Ellisman, M.H. (2006)
Transform-based backprojection for volume reconstruction of large
format electron microscope tilt series. J. Struct. Biol., 154, 144–167.
Human Molecular Genetics, 2012, Vol. 21, No. 4899
at University of California, San Diego on October 10, 2012
http://hmg.oxfordjournals.org/
Downloaded from