Increasing microtubule acetylation rescues axonal transport and locomotor deficits caused by LRRK2 Roc-COR domain mutations

Leucine-rich repeat kinase 2 (LRRK2) mutations are the most common genetic cause of Parkinson’s disease. LRRK2 is a multifunctional protein affecting many cellular processes and has been described to bind microtubules. Defective microtubule-based axonal transport is hypothesized to contribute to Parkinson’s disease, but whether LRRK2 mutations affect this process to mediate pathogenesis is not known. Here we find that LRRK2 containing pathogenic Roc-COR domain mutations (R1441C, Y1699C) preferentially associates with deacetylated microtubules, and inhibits axonal transport in primary neurons and in Drosophila, causing locomotor deficits in vivo. In vitro, increasing microtubule acetylation using deacetylase inhibitors or the tubulin acetylase αTAT1 prevents association of mutant LRRK2 with microtubules, and the deacetylase inhibitor trichostatin A (TSA) restores axonal transport. In vivo knockdown of the deacetylases HDAC6 and Sirt2, or administration of TSA rescues both axonal transport and locomotor behavior. Thus, this study reveals a pathogenic mechanism and a potential intervention for Parkinson’s disease.

D ominant mutations in leucine-rich repeat kinase 2 (LRRK2) are the most common genetic cause of Parkinson's disease (PD), and genome-wide association studies have highlighted the LRRK2 locus as a risk factor for sporadic PD [1][2][3][4][5] . LRRK2 is a multifunctional, multi-domain protein and dominant mutations have been identified in the Ras of complex (Roc) GTPase protein domain (R1441C, R1441G, R1441H), the carboxy-terminal of Roc (COR) domain (Y1699C) and the kinase domain (G2019S and I2020T). Mutations in the Roc and COR domains decrease GTPase activity 6,7 , whereas the G2019S mutation increases kinase activity 6,8,9 . How LRRK2 mutations confer toxicity is not yet clear, but may involve the possible physiological function of LRRK2 in the maintenance of axonal integrity. Indeed, there is evidence that the overexpression of LRRK2 causes neurite shortening, whereas the loss of LRRK2 function in mouse neurons results in elongation of neuronal processes and increased branching [10][11][12][13] . The molecular details underlying mutant LRRK2's effects on neurites are not fully understood, but may involve interactions with microtubules [14][15][16][17][18] . The architecture of neurons makes them particularly dependent on efficient intracellular transport of cargoes such as organelles and proteins to maintain structural integrity and function. Since defective axonal transport, which utilizes microtubules as transport tracks, has been proposed as a mechanism for PD [19][20][21][22][23][24] , we hypothesized that mutant LRRK2 may disturb microtubule-based axonal transport.
Here we show that LRRK2 Roc-COR mutants disrupt axonal transport in vitro and in vivo to induce locomotor deficits by binding to deacetylated microtubules. Increasing microtubule acetylation prevents microtubule association, restores axonal transport and rescues locomotor deficits caused by LRRK2 Roc-COR mutants. Thus, our data reveals a pathogenic mechanism and a potential therapeutic intervention for PD.

Roc-COR mutations inhibit axonal transport and locomotion.
To study the effect of pathogenic LRRK2 mutations on axonal transport, we expressed wild-type (WT) or mutant (R1441C, Y1699C or G2019S) LRRK2 variants in rat cortical neuron cultures, and analysed the movement of mitochondria from time-lapse recordings. Whereas the expression of LRRK2-WT or LRRK2-G2019S did not affect mitochondrial transport, both LRRK2-R1441C and LRRK2-Y1699C significantly inhibited the transport of mitochondria by disrupting both anterograde and retrograde motility (Fig. 1a). Detailed analysis revealed that both LRRK2-R1441C and LRRK2-Y1699C decreased anterograde and retrograde transport frequencies, with a concomitant increase in the duration of pauses and decreased velocities ( Supplementary  Fig. 1).
To investigate if this axonal transport defect occurred in vivo, we analysed the transport of mitochondria in motor neurons of Drosophila larvae expressing human LRRK2 variants. Similar to mammalian neurons, the expression of LRRK2-WT or LRRK2-G2019S did not adversely affect mitochondrial transport, whereas the LRRK2-R1441C or LRRK2-Y1699C mutants reduced the number of motile mitochondria in both anterograde and retrograde directions (Fig. 1b). Moreover, the effects are evolutionarily conserved as the expression of Drosophila Lrrk-R1069C (equivalent to LRRK2-R1441C) or Lrrk-Y1383C (equivalent to LRRK2-Y1699C) inhibited both anterograde and retrograde transport, while the expression of Lrrk-WT similarly had no effect on transport ( Supplementary Fig. 2). Together these results demonstrate that the expression of pathogenic Roc-COR variants of LRRK2 or its Drosophila homologue Lrrk inhibit bidirectional axonal transport of mitochondria in vitro and in vivo.
The Drosophila model enabled us to investigate whether the aberrant axonal transport resulted in a functional deficit of the motor system. Adult flies expressing variant forms of LRRK2 or Lrrk in motor neurons were assayed for climbing and flight ability. Flies expressing WT LRRK2/Lrrk or LRRK2-G2019S showed no defect in locomotion compared with controls, consistent with a lack of effect on axonal transport ( Fig. 1c and Supplementary Fig. 2). However, LRRK2-R1441C and LRRK2-Y1699C caused a significant decrease in both climbing and flight ability (Fig. 1c). Consistent with a conserved effect of these mutations, Lrrk-R1069C and Lrrk-Y1383C also caused locomotor deficits ( Supplementary Fig. 2).
Roc-COR mutants form filamentous structures on microtubules. The bidirectional nature of the axonal transport defect suggested the disturbance of a common transport component such as microtubules. It has been reported that LRRK2 can interact with microtubules and even alter tubulin acetylation [14][15][16][17][18] . Importantly, the acetylation status of microtubules is known to influence the rate of transport along them, with increased acetylation promoting axonal transport 25,26 . Therefore, we hypothesized that LRRK2 Roc-COR mutants may affect the equilibrium of microtubules. To address this, we analysed the association of LRRK2 variants with microtubules in mammalian CV1 and HEK293 cells, focusing specifically on acetylated microtubules. In contrast to LRRK2-WT, we found that LRRK2-R1441C and LRRK2-Y1699C had a propensity to form filamentous structures that appeared to align with microtubules (Fig. 2). These structures are reminiscent of previously reported skein-like LRRK2 structures 17,27 . Furthermore, where LRRK2-R1441C or LRRK2-Y1699C was present, a-tubulin acetylation was lacking. These observations are consistent with LRRK2-R1441C and LRRK2-Y1699C inhibiting axonal transport by interfering with microtubule acetylation.
Increasing microtubule acetylation inhibits LRRK2 filaments. We next investigated whether the interaction of LRRK2 variants with deacetylated microtubules was reversible and sensitive to microtubule acetylation. We treated CV1 and HEK293 cells with the broad deacetylase inhibitor trichostatin A (TSA), which promotes a-tubulin acetylation 26 , and analysed filament formation. In both CV1 and HEK293 cells, TSA treatment significantly reduced the formation of filamentous LRRK2 (Fig. 2) The family of histone deacetylases (HDACs) is classified into four groups based on function and DNA sequence similarity. Classes I, IIa and IIb are considered the classical HDACs whose activities are inhibited by TSA. Class III comprises a family of NAD þ -dependent proteins called the Sirtuins (SIRTs) that are not affected by TSA. Class IV is an atypical class based solely on DNA sequence similarity to the others. Both the class IIb member HDAC6 and class III member SIRT2 have been shown to deacetylate a-tubulin 28,29 and conversely their inhibition increases microtubule acetylation. Thus, although in our assays HDAC6 was likely to be the main target of TSA, other pathways could have been affected by TSA treatment. To more selectively target microtubule acetylation, we treated HEK293 cells expressing LRRK2 Roc-COR mutants with the specific HDAC6 inhibitor tubastatin-A [30][31][32] or the SIRT2-specific inhibitor AGK2 (ref. 33). Specific inhibition of either HDAC6 or SIRT2 prevented the formation of filamentous LRRK2 to the same extend as TSA (Fig. 3).
To further characterize the potential role of microtubule acetylation in the LRRK2 Roc-COR mutant filamentous phenotype, we expressed the a-tubulin acetyl transferase aTAT1 (also called MEC17) 34,35 . Because a-tubulin is the major cellular ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6245 substrate of aTAT1, the overexpression of aTAT1 specifically increases microtubule acetylation 36 . aTAT1 expression also significantly reversed the filamentous LRRK2 Roc-COR mutant phenotype (Fig. 4) confirming that microtubule acetylation regulates the association of LRRK2 Roc-COR mutants with microtubules.
Together these findings support previous reports that LRRK2 associates with microtubules and indicate that LRRK2 Roc-COR mutants interfere with a-tubulin acetylation, which can be restored by the inhibition of the deacetylase enzymes HDAC6 and SIRT2 or by the overexpression of aTAT1.
TSA restores axonal transport and locomotion. Reversal of the association of LRRK2 Roc-COR mutants with deacetylated microtubules prompted us to address whether modulating microtubule acetylation could also reverse the axonal transport defects in vitro and in vivo. We first targeted the tubulin deacetylases in vivo using transgenic Drosophila HDAC6 or Sirt2 RNA interference (RNAi) lines. Loss of either HDAC6 or Sirt2 effectively restored axonal transport of mitochondria in LRRK2 Roc-COR mutant transgenic flies (Fig. 5a,b). Moreover, the knockdown of HDAC6 or Sirt2 was able to significantly rescue climbing and flight deficits caused by mutant LRRK2 (Fig. 5c).
The ability of TSA to prevent mutant LRRK2 from aberrant binding to microtubules makes this an attractive therapeutic intervention. We tested this idea by treating primary rat neurons expressing LRRK2-Y1699C with TSA. Encouragingly, this was able to restore axonal transport to control levels ( Fig. 6a). To extend this observation in vivo, Drosophila expressing the LRRK2 variants were raised on TSA-dosed food. Remarkably, this completely rescued the transport defects caused by LRRK2-R1441C or Y1699C (Fig. 6b). Similarly, transport defects caused by Lrrk-R1069C or Y1383C expression were also restored by feeding TSA (Supplementary Fig. 3). Thus, both genetic and pharmacological approaches in vitro and in vivo show that increasing microtubule acetylation can rescue LRRK2 Roc-COR mutant-induced axonal transport defects.
Finally, we asked whether treatment with TSA would be able to rescue the locomotor deficits cause by the defective axonal transport even after the onset of the phenotype. We took newly emerged flies expressing WT and mutant forms of LRRK2 or Lrrk that had been raised on normal food, which as we previously showed already exhibited climbing and flight deficits (Fig. 1c). These young adult flies were then fed TSA for 5 days and tested for locomotion. Notably, even with this modest increase in age, the locomotor deficit had worsened in control-fed animals ( Fig. 6c and Supplementary Fig. 3c). However, the administration of TSA was sufficient to fully restore locomotor ability in the ARTICLE LRRK2-R1441C-or Y1699C-and Lrkk-R1069C-or Y1383Cexpressing animals ( Fig. 6c and Supplementary Fig. 3c). These results demonstrate that increasing microtubule acetylation either genetically by knockdown of deacetylase enzymes or by systemic administration of a deacetylase inhibitor is sufficient to restore a primary cellular defect and reverse the associated behavioural disruption, even after the onset of behavioural symptoms. Discussion Disruption of axonal transport is an early feature in a number of neurodegenerative diseases including PD 19 . Analysis of axonal transport of WT and pathogenic mutants of a-synuclein have revealed reduced transport rates of the mutants, which may contribute to the formation of Lewy bodies 21 . PINK1 and Parkin, mutations in which cause the majority of early-onset recessive familial PD 37,38 , have been shown to directly affect mitochondrial transport 23,24 and this in turn regulates mitophagy 39 . Halting of axonal transport has also been observed in mitochondrial neurotoxin models of sporadic PD [40][41][42] . ARTICLE In this study, we investigated whether PD-associated mutations in LRRK2 affect axonal transport and cause dysfunction of the motor system. We found that pathogenic mutations in the Roc-COR domain of LRRK2 (R1441C and Y1699C) disrupt axonal transport of mitochondria in vitro in cultured neurons and in vivo, and cause locomotor deficits in Drosophila. The effects are evolutionarily conserved since the expression of mutant Drosophila Lrrk variants also inhibited transport and motor ability.
LRRK2 has been shown to interact with microtubules but the physiological relevance of this interaction is poorly understood [14][15][16][17][18] . We found that both LRRK2-R1441C and Y1699C mutants, but not LRRK2-WT, formed filament-like structures that decorated deacetylated microtubules. Previous work has shown that acetylation of microtubules regulates axonal transport 25,26 and has implicated LRRK2 in the regulation of microtubule acetylation 18 . We therefore increased microtubule acetylation using pharmacological inhibitors or RNAi knockdown of the deacetylase enzymes HDAC6 and SIRT2, or by the expression of aTAT1 and found that this prevented the decoration of microtubules by LRRK2-R1441C and Y1699C. Importantly, increasing microtubule acetylation restored axonal transport to WT levels in vitro and in vivo.
Interestingly the kinase domain mutant LRRK2-G2019S did not affect axonal transport. Several groups have shown that the R1441C and Y1699C mutations inhibit the GTPase activity of LRRK2 but have normal kinase activity 6,7,43,44 , whereas LRRK2-G2019S has increased kinase activity but normal GTPase activity compared with WT LRRK2 (refs 8,9,45,46). Thus, our results suggest that the inhibition of axonal transport is likely linked to decreased LRRK2 GTPase activity. This is in line with increasing evidence suggesting that the GTPase activity of LRRK2 plays an important role in pathogenesis 47 . Interestingly, the pharmacological inhibition of LRRK2 kinase activity has been shown to induce a redistribution of LRRK2 that appears similar to the microtubule-attached distribution we observed with LRRK2 Roc-COR mutants 48 and the binding of LRRK2 to tubulin was significantly enhanced by LRRK2 kinase inhibition in an indirect reporter assay 16 . However, since the LRRK2 Roc-COR mutants analysed here have normal kinase activity 6,7,43,44 , our data suggest that the GTPase activity of LRRK2 plays a key role in LRRK2's interaction with microtubules. The effect of LRRK2 kinase inhibitors may be explained by a model in which LRRK2 kinase regulates the GTPase via phosphorylation 47 .
The Drosophila model allowed us to extend our observations of defective axonal transport in a number of ways. First, it allowed us to address whether the inhibition of axonal transport by LRRK2 Roc-COR mutants had any physiological consequences in adult motor behaviours. We found that Drosophila expressing LRRK2 Roc-COR mutants also resulted in a deficit of the motor system, causing a decrease in both climbing and flight ability. Furthermore, when we restored axonal transport by increasing microtubule acetylation using HDAC6 or Sirt2 RNAi, this also restored normal locomotor ability. Thus, these results argue that inhibition of axonal transport can cause functional deficits in vivo. Second, this system also enabled us to assess whether pharmacological restoration of acetylated microtubules could reverse the locomotor phenotype after the onset of locomotor symptoms. Indeed, we found that the systemic administration of TSA was able to fully restore locomotor ability, showing that the effects of the LRRK2 Roc-COR mutants are reversible and, at least  in the fly, are amenable to post hoc treatment. Thus, our study adds PD to the growing list of neurodegenerative diseases that involve disturbed axonal transport in which deacetylase inhibitors have been shown to confer protection including Charcot-Marie-Tooth disease, Alzheimer's disease, amyotrophic lateral sclerosis and Huntington's disease 26,[49][50][51][52][53][54][55] .
In conclusion, our results reveal that mutant LRRK2-associated PD involves defective axonal transport and suggest that treating with deacetylase inhibitors to restore transport may have therapeutic potential.
Cell culture and plasmid transfection. CV1 and HEK293 cells obtained from the ATCC were maintained in Dulbecco's modified Eagle's medium (Invitrogen) containing 4.5 g l À 1 glucose, 10% foetal bovine serum (Sera Laboratories), 2 mM L-glutamine (Invitrogen) and 1 mM sodium pyruvate (Sigma) and were transfected using Exgen500 (Fermentas) or Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. All cells were used in experiments 16-24 h post transfection.
Cortical neurons were isolated and transfected as previously described [56][57][58][59] . Briefly, cortical neurons were isolated from embryonic day 18 rat embryos and cultured on glass coverslips coated with poly-L-lysine in six or 12-well plates in neurobasal medium containing B27 supplement (Invitrogen), 100 IU ml À 1 penicillin, 100 mg ml À 1 streptomycin and 2 mM L-glutamine. Cells were cultured for 7 days and then transfected using a calcium phosphate Profection kit (Promega). For studies on the effects of LRRK2 on axonal transport, neurons were transfected with twice the amount of LRRK2 plasmid compared with DsRed-Mito to ensure that all cells visualized for mitochondrial transport expressed LRRK2 and this was confirmed by AcGFP1 fluorescence.
Immunofluorescence analysis of LRRK2 microtubule interaction. Immunostaining was performed as described previously 60 . Briefly, CV1 or HEK293 cells on glass coverslips were fixed with 3.7% formaldehyde in PBS for 20 min at room temperature. After washing with PBS, residual formaldehyde was quenched by incubation with 50 mM NH 4 Cl in PBS for 15 min at room temperature, followed by a second round of washing with PBS. Subsequently, the cells were permeabilized with À 20°C MeOH for 20 min or by incubation with 0.2% Triton-X-100 in PBS for 3 min. MeOH or Triton-X-100 was removed by washing with PBS.
After fixing, the cells were incubated with PBS containing 0.2% fish gelatine (PBS/F) for 30 min at room temperature and then with the primary antibody in PBS/F for 1 h. After washing with PBS/F, the cells were incubated with secondary antibody in PBS/F for 45 min at room temperature. After a final wash, the samples were mounted in Moviol (Calbiochem) containing 1% DABCO or Dako fluorescence mounting medium.
Confocal images were recorded with an LSM510Meta (Carl Zeiss) or a Leica SP5 (Leica Microsystems) confocal microscope using a Â 63, 1.4NA objective.
Axonal transport of mitochondria in rat cortical neurons. Live microscopy of mitochondrial axonal transport in cortical neurons was performed with an Axiovert S100 microscope (Carl Zeiss) equipped with a Lambda LS Xenon-Arc light source (Sutter Instrument Company, Novato, CA), an EGFP/DsRed filterset (Chroma Technology Corp.), Â 40 EC Plan-Neofluar 1.3 N.A. objective (Zeiss), Lambda 10-3 filter wheel (Sutter Instrument Co.) and a Photometrics Cascade-II 512B High Speed EMCCD camera (Photometrics). Post transfection (36-48 h), neurons on coverslips were transferred to a custom observation chamber mounted on the stage of the microscope. The cells were maintained at 37°C using an objective heater (Tempcontrol 37-2, Carl Zeiss) and 'The Box' Microscope Temperature Control System (Life Imaging Systems). Mitochondrial movements were recorded for 10 min with 3 s time-lapse interval using Metamorph software (Molecular Devices).
Image analysis was performed with ImageJ developed by Wayne Rasband (NIH, Bethesda, USA; http://rsb.info.nih.gov/ij/) extended with custom plug-ins, or Metamorph. Calculations of mitochondrial transport parameters were as described before 57,59 except that the overall transport of mitochondria in Figs 1 and 5 were analysed from kymographs using the SlopeToVelocity plugin described previously 59 . Briefly, this plugin calculates the velocity by measuring the distance between the position of individual mitochondria at the start and end of time-lapse recordings and dividing by the time elapsed. This yields an overall velocity of transport that includes anterograde and retrograde movements and stationary periods. Mitochondria were subsequently classified as motile (velocity 40.1 mm s À 1 ) or stationary (velocity r0.1 mm s À 1 ).  65 . Flight and climbing assays were performed as previously described 66 .
Dissection and imaging of axonal transport in Drosophila. Live dissections of wandering third instar larvae were performed as previously described 67 . Individual larvae were placed on Sylgard slide and pinned on both ends with the dorsal side up. Larval body wall muscles were carefully opened using microdissection scissors. Internal organs were removed with forceps without disturbing the ventral ganglion and motor neurons. The cuticle was laterally expanded on both sides and pinned. The whole preparation was washed in 1 ml dissecting buffer containing 128 mM NaCl, 1 mM EGTA, 4 mM MgCl 2 , 2 mM KCl, 5 mM HEPES and 36 mM sucrose, pH 7.2. The buffer was replaced and the larvae kept immersed in buffer during imaging. Imaging was done on an Olympus FV1000 confocal using Â 60 water immersion objective (NA 0.90, Olympus LUMPLFL). Time-lapse movies were acquired at the rate of 1 frame per 5 s for 500 s. A minimum of 10 movies (one movie per larva) were taken for each genotype/treatment.
Analysis of mitochondrial movement in Drosophila. All movies were analysed using ImageJ software. The movies were processed in ImageJ and kymographs were generated as previously described 59 . Using kymographs, the number of mitochondria moving in anterograde and retrograde directions were quantified along with the number of stationary mitochondria. Mitochondria were scored as moving if the organelle is in motion for at least 60% of the movie. All scoring was performed blind to the experimental condition.
Drug treatment in Drosophila. Crosses were raised in standard medium supplemented with 10 mM TSA (Sigma-Aldrich T8552). Dimethylsulphoxide (0.06%) was used as a control. For adult climbing, crosses with D42-GAL4 were raised in standard medium and young progeny of 1-2 days were collected. Flies were starved in empty vials for 3-4 h at 25°C and then placed in vials with filter papers soaked with 1 ml solution containing 10% sucrose and 100 mM TSA 68 . Dimethylsulphoxide (0.6%) was used as a control. The filter papers were checked every day and replaced after 48 h. Climbing assays were performed after 5 days, that is, animals were 6-7 days old.
Statistical analysis. Calculations and statistical analysis were performed using Excel (Microsoft Corporation, Redmond, WA), and Prism software (GraphPad Software, Inc., San Diego, CA). Statistical significance was determined by one-way or two-way analysis of variance (ANOVA) followed by Fisher's least significant difference post hoc test or by t-test as indicated in the figure legends.