Endogenous Rab29 does not impact basal or stimulated LRRK2 pathway activity

Mutations that enhance LRRK2 protein kinase activity cause inherited Parkinson's disease. LRRK2 phosphorylates a group of Rab GTPase proteins, including Rab10 and Rab12, within the effector-binding switch-II motif. Previous work has indicated that the PARK16 locus, which harbors the gene encoding for Rab29, is involved in Parkinson's, and that Rab29 operates in a common pathway with LRRK2. Co-expression of Rab29 and LRRK2 stimulates LRRK2 activity by recruiting LRRK2 to the surface of the trans Golgi network. Here, we report that knock-out of Rab29 does not influence endogenous LRRK2 activity, based on the assessment of Rab10 and Rab12 phosphorylation, in wild-type LRRK2, LRRK2[R1441C] or VPS35[D620N] knock-in mouse tissues and primary cell lines, including brain extracts and embryonic fibroblasts. We find that in brain extracts, Rab12 phosphorylation is more robustly impacted by LRRK2 inhibitors and pathogenic mutations than Rab10 phosphorylation. Transgenic overexpression of Rab29 in a mouse model was also insufficient to stimulate basal LRRK2 activity. We observed that stimulation of Rab10 and Rab12 phosphorylation induced by agents that stress the endolysosomal system (nigericin, monensin, chloroquine and LLOMe) is suppressed by LRRK2 inhibitors but not blocked in Rab29 deficient cells. From the agents tested, nigericin induced the greatest increase in Rab10 and Rab12 phosphorylation (5 to 9-fold). Our findings indicate that basal, pathogenic, as well as nigericin and monensin stimulated LRRK2 pathway activity is not controlled by Rab29. Further work is required to establish how LRRK2 activity is regulated, and whether other Rab proteins can control LRRK2 by targeting it to diverse membranes.


Introduction
Autosomal dominant missense mutations that hyperactivate LRRK2 (leucine-rich repeat kinase 2) are one of the most common causes of familial Parkinson's disease (PD) [1][2][3][4]. Age of onset and progression of LRRK2-driven PD is virtually indistinguishable from sporadic PD, which comprises the vast majority of the patient population. LRRK2 is a large, multi-functional protein kinase that encodes two central catalytic regions, a Roc-type GTPase domain adjacent to a COR (C-terminal of Roc) domain, which is followed by a serine/threonine protein kinase domain. These enzymatic regions are surrounded by several domains, including N-terminal armadillo and ankyrin domains, leucine-rich repeats and a C-terminal WD-40 repeat [5]. The most prevalent LRRK2 pathogenic variants map to either the GTPase Roc [N1437H, R1441C/G/H] or COR [Y1699C] domains, or the kinase domain [G2019S, I2020T], and act as gain-of-function mutations by enhancing LRRK2 kinase activity [6][7][8][9][10]. The mutations within the GTPase Roc/COR domain are proposed to inhibit GTPase activity and enhance GTP binding [11][12][13]. The mutations within the GTPase domain do not directly activate LRRK2 activity in vitro, but enhance interaction with Rab29 located at the Golgi [14,15]. This leads to the recruitment of LRRK2 to the Golgi membrane surface which enhances its kinase activity, as evidenced by increased autophosphorylation at Ser1292 and phosphorylation of its physiological Rab protein substrates, through a yet undefined mechanism [14,16,17]. Mutations within the kinase domain directly stimulate LRRK2 activity by promoting a more closed, active conformation of the catalytic moiety [18][19][20]. The VPS35 [D620N] autosomal dominant mutation that causes PD markedly elevates Rab protein phosphorylation by LRRK2 through an unknown mechanism [21]. LRRK2 is constitutively phosphorylated on several well-studied serine residues in its N-terminus, specifically Ser910, Ser935, Ser955 and Ser973, and these sites are rapidly dephosphorylated upon pharmacological inhibition of LRRK2 [22,23]. LRRK2 protein kinase inhibitors are currently in early stage clinical trials for LRRK2-driven PD [24,25].
Well-characterized and validated substrates of LRRK2 comprise a subset of Rab GTPases that include Rab8A, Rab10, Rab12 and Rab29 [7,26]. Rab GTPases are crucial regulators of intracellular vesicle trafficking, implicated in vesicle formation and transport between target membranes in a tightly controlled network [27]. They influence biology by interacting with specific effector proteins when complexed to GTP. LRRK2 phosphorylates Rab proteins at a highly conserved Ser/Thr residue located at the center of the effector-binding region of these enzymes that is also known as the Switch-II motif [7,26,28]. This phosphorylation event appears to act in two ways. Firstly, it prevents Rab proteins interacting with many of their known interactors including guanine nucleotide exchange factors (GEFs) and guanine nucleotide dissociation inhibitors (GDIs) that are required for the shuttling of Rab proteins between membrane compartments. This results in the LRRK2-phosphorylated Rab proteins accumulating on the surface of the compartment on which they are phosphorylated [7,29]. Secondly, LRRK2-phosphorylated Rab8A and Rab10 bind preferentially to a set of effectors, such as RILPL1 and RILPL2, which are implicated in ciliogenesis [26]. These effectors possess a RH2 domain that functions as a phospho-Rab recognition domain [30]. Pathogenic LRRK2 mutants decrease primary cilia formation in cell culture in a manner that is rescued upon LRRK2 inhibition [26]. The ability of LRRK2 to inhibit ciliogenesis requires RILPL1 binding to LRRK2-phosphorylated Rab8A and Rab10 [31,32]. Recent work showed that LRRK2-phosphorylated Rab proteins are dephosphorylated by a highly selective PPM1H protein phosphatase [33].
Significant effort has focused on Rab29, also known as Rab7L1, and its possible roles in regulating LRRK2. The gene encoding Rab29 lies within a genetically complex locus termed PARK16, which is implicated with increased PD risk [34][35][36][37]. The PARK16 locus contains five genes, and it is not clear which of these genes is relevant for PD or how the numerous variants identified within this locus affect gene expression and/or function. Single nucleotide polymorphisms in non-coding regions of the PARK16 locus have been linked to increasing the transcriptional regulation of Rab29 mRNA [15,38,39]. Several earlier studies alluded to the possibility of Rab29 and LRRK2 acting in converging pathways, by demonstrating epistatic interactions between polymorphisms in the LRRK2 and Rab29 genes that increase PD risk [39,40]. Physical interaction between LRRK2 and Rab29, either in vitro or based on a co-immunoprecipitation analysis, has also been demonstrated [15,39,41,42]. Furthermore, analysis of genetic models reveals that Rab29 and LRRK2 operate co-ordinately to control axon elongation in Caenorhabditis elegans, and lysosomal trafficking and kidney pathology in mice [43]. A recent study reports that combined knock-out of LRRK2 and Rab29 does not result in a PD-relevant neuronal pathology or behavioral abnormalities [44]. Rab29 has been implicated in maintaining Golgi morphology and in mediating the retrograde trafficking of the mannose-6-phosphate receptor (M6PR), which recognizes and delivers lysosomal enzymes from the trans Golgi to late endosomes and lysosomes [45,46]. An intriguing finding that implicates Rab29 in immune response demonstrated its recruitment to Salmonella typhi-containing vacuoles and Rab29 involvement in the generation of typhoid toxin transport intermediates that release the toxin into the extracellular environment [47].
Rab29 belongs to a subfamily of Rab GTPases with Rab32 and Rab38, which are localized to melanosomes and are involved in regulating the trafficking of melanogenic enzymes between the trans Golgi and melanosomes [48]. Rab29 is unique among the Rab proteins targeted by LRRK2 in that it possesses two adjacent phosphorylated residues within its Switch-II motif, namely Thr71 and Ser72. Ser72 aligns with the phosphorylation site found in other LRRK2 substrates. To our knowledge there is no evidence that endogenous Rab29 is directly phosphorylated by LRRK2, however, in overexpression studies, LRRK2 triggers phosphorylation of both Thr71 and Ser72 in a manner that is blocked with LRRK2 inhibitors [14,26]. Based on mutagenesis overexpression experiments, phosphorylation of these sites on Rab29 was proposed to function as a negative feedback loop to block the activation of LRRK2 by Rab29 [14]. Rab32 and Rab38 are not phosphorylated by LRRK2 and do not possess a Ser/Thr residue at the equivalent position within their switch-II motif [26]. Recent work has established that the activation of LRRK2 by Rab29 occurs irrespective of the identity of the membrane to which Rab29 is attached [29], and requires Rab prenylation and nucleotide binding by both LRRK2 and Rab29 [16,29]. Thus, numerous genetic studies in addition to data presented in vitro, in cells, and in model organisms, provide substantial evidence that LRRK2 and Rab29 pathways intersect.
Previous work showing that LRRK2 is activated by recruitment to the Golgi via interaction with Rab29 is largely based in overexpression experiments in which both Rab29 and wildtype or pathogenic mutants of LRRK2 are co-expressed. In this study, we sought to investigate the physiological relevance of Rab29 as a regulator of endogenous LRRK2 in cell lines as well as in mouse tissues. We describe our efforts to thoroughly characterize four different mouse models that we generated to assess Rab29 impact on basal activity of wildtype and the LRRK2[R1441C] pathogenic mutant. Our data demonstrate that knock-out or moderate transgenic overexpression of Rab29 in a new mouse model we created, does not significantly impact the ability of LRRK2 to phosphorylate Rab10 or Rab12. Furthermore, we show that knock-out of Rab29 does not impact the ability of the VPS35[D620N] mutation or monovalent cation ionophore antibiotics nigericin and monensin to promote Rab10 and Rab12 protein phosphorylation. Our data indicate that Rab29 is not a major regulator of basal wildtype or LRRK2[R1441C] activity as measured in whole cell or tissue extracts that we have analyzed. Further work is, therefore, required to clarify how LRRK2 activity is regulated.

Generation of MJFF rabbit monoclonal Rab29 total antibodies (MJF-30)
Rabbit immunization and rabbit antibody generation was performed by Abcam Inc. (Burlingame, CA). To generate the Rab29 total antibodies, full length recombinant proteins as well as N-terminal and C-terminal peptides were used. Abcam performed three subcutaneous injections using the immunogens conjugated with keyhole limpet hemocyanin (KLH), followed by two subcutaneous injections using the immunogens conjugated with ovalbumin. Target immunogen is described in Table 1. Following the initial injections with full-length proteins, booster immunizations using N-terminal and C-terminal peptides were performed. The provided bleeds (one bleed pre-immunization, two bleeds after immunization with full-length protein, one bleed after additional immunization with the N-and C-terminal peptides) from immunized animals (six rabbits -E8767-E8772) were tested at 1 : 1000 dilution using lysates of A549 wildtype and Rab29 knock-out cell lysates. Rabbits producing the best antibody were chosen for monoclonal antibody generation. Hybridoma fusion was performed according to an established protocol [50]. Abcam used a process omitting the multi-clone stage and provided 81 single clones directly. Single-clone supernatants were screened using immunoblots of A549 wildtype and Rab29 knock-out cell lysates. The top eight clones that gave the most robust signal were additionally tested using MEF lysates to check for human/mouse Rab29 specificity. Clone #124 was able to detect both human and mouse Rab29, while clone #104, which only detected human Rab29, showed the strongest and cleanest signal. Based on these results, clones #104 (human selective catalog number ab256527) and clone #124

Mice
Mice selected for this study were maintained under specific pathogen-free conditions at the University of Dundee (U.K.). All animal studies were ethically reviewed and carried out in accordance with the Animals   Figure S2 (transgenic Rab29 overexpression model), mice ranging from 3 to 4 months and of the indicated genotypes were dosed, killed and the tissues were isolated as outlined above. For the studies involving 18-24 mice which took place over multiple days, the genotypes and treatments were randomized each day to account for any temporal differences, such as light and ventilation. A general overview of the mouse models utilized in this study is outlined in Table 2.

Generation of Rab29 knock-out mice
Rab29 knock-out mice are made available through the Wellcome Trust Sanger Institute, distributed by Infrafrontier: EMMA mouse repository (EM: 05517), and were characterized previously [43]. LacZ-knock-in Rab29 tm1a(EUCOMM)Wtsi mice were bred with Taconic Total Body Cre mice expressing Cre recombinase (Model 12524), which recognizes the loxP sites that flank the inserted promoter-driven neomycin cassette and exon 4 of Rab29, which is critical for expression. Following the deletion of exon 4, the mice were then bred and maintained on a C57Bl/6j background to remove the Cre recombinase allele, and to produce the experimental animals used in this study. The genotypes of the Rab29 knock-out mice were confirmed by PCR using genomic DNA isolated from ear biopsies and primers that amplify the entire Rab29 gene, as well as by immunoblotting.

Preparation of mouse tissue lysates
Mouse tissues were collected and snap frozen in liquid nitrogen. Snap frozen tissues were weighed and quickly thawed on ice in a 10-fold volume excess of ice-cold lysis buffer containing 50 mM Tris-HCl pH 7.4, 1 mM EGTA, 10 mM 2-glycerophosphate, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 270 mM sucrose, supplemented with 1 μg/ml microcystin-LR, 1 mM sodium orthovanadate, complete EDTA-free protease inhibitor cocktail (Roche), and 1% (v/v) Triton X-100. Tissue was homogenized using a POLYTRON homogenizer (KINEMATICA), employing three rounds of 10 s homogenization with 10 s intervals on ice. Lysates were centrifuged at 20 800g for 30 min at 4°C and supernatant was collected for subsequent Bradford assay and immunoblot analysis.

Quantitative real-time RT-PCR analysis of Rab29 mRNA from mouse tissue
Frozen tissue samples were ground to a fine powder in a vessel submerged in liquid nitrogen using a mallet. Powdered tissue (∼5 mg) was dissolved in lysis buffer provided by RNeasy micro kit (Qiagen), supplemented with 1% (v/v) β-mercaptoethanol. Tissue was homogenized in lysis buffer using IKA VIBRAX VXR basic orbital shaker for 5 min at 4°C at 1000 rpm and total RNA was extracted from tissue lysate following RNeasy micro kit instructions. cDNA was synthesized from total RNA extracts using Bio-Rad iScript cDNA synthesis kit (#170-8891), using a starting template of 150 ng total RNA. The Bio-Rad Sso EvaGreen Supermix (#1725201) was used to set up qPCR reactions in a 384-well plate format according to the manufacturer's instructions and 20 μl reactions were prepared in duplicate. Real-time quantitative PCR primers for mouse Rab29 were designed using NCBI Primer Blast and the sequences are as follows: 5 0 -AGGCCATGAGAGTCCT-CGTT-3 0 (forward) and 5 0 -GGGCTTGGCTTGGAGATTTGA-3 0 (reverse). The β-actin internal control primers employed for real-time quantitative PCR analysis are described in [52]: 5 0 -CACTATCGGCAATGAGCGGTTC-C-3 0 (forward) and 5 0 -CAGCACTGTGTTGGCATAGAGGTC-3 0 (reverse). Primers were ordered from Sigma as lyophilized and reconstituted in Milli-Q water to a stock concentration of 100 μM and further diluted to a 10 μM working stock. The PCR efficiency of the primers was validated using the relative standard curve method. The qPCR reactions were run using the Bio-Rad CFX384 Real-Time System C1000 Thermal Cycler, and raw data were collected for duplicate reactions from four biological replicates per genotype for three different tissues. The relative quantification of Rab29 and β-actin mRNA was undertaken using the comparative Ct (cycle threshold) method, which employs the formula RQ = 2 (−ΔΔCt) [53].
LRRK2 fragment purification and Rab29 pulldown assay A total of 1.5 mg of tissue lysate ( prepared as described above) was diluted to 10 ml using low salt buffer to reduce Triton X-100 concentration to ∼0.03% and supplemented with GTP-γ-S at a final concentration of 100 μM. Tissue lysate was incubated with 200 μl nickel-NTA agarose with immobilized 6-His LRRK2 1-552 or 6-His for 2 h rotating at 4°C. The resin was washed three times with 5 ml low salt buffer. An equal volume of elution buffer containing 500 mM Imidazole and low salt buffer was added to the resin and the slurry was transferred to Corning Costar spin-X centrifuge tube filters (Corning #8161). The resin was incubated with elution buffer for 20 min rotating at 4°C. Protein was eluted in spin-X tubes by centrifugation at 500g for 5 min at 4°C. Protein was concentrated to 80 μl using Amicon Ultra centrifugal filters, 10 K (#UFC901024). 15% of the eluted protein was analyzed by immunoblotting.

Microsome enrichment by fractionation
Fractionation to detect LRRK2 auto-phosphorylation at Ser1292 in mouse tissues was described previously [54]. Briefly, snap frozen mouse lungs were homogenized on ice in 15% (w/v) sedimentation buffer containing 3 mM Tris-HCl pH 7.4, 250 mM sucrose, 0.5 mM EGTA, 10 mM 2-glycerophosphate, 50 mM NaF, 5 mM sodium pyrophosphate, 1 μg/ml microcystin-LR, 1 mM sodium orthovanadate and complete EDTA-free protease inhibitor cocktail (Roche). Homogenates were centrifuged at 3000g for 10 min at 4°C, and the supernatant was centrifuged again at 5000g for 10 min at 4°C to produce cleared homogenate. Two-hundred microliters of cleared homogenate was supplemented with 10X lysis buffer (200 mM Tris-HCl pH 7.4, 1.5 M NaCl, 10 mM EGTA, 25 mM sodium pyrophosphate, 10 mM 2-glycerophosphate, 10 mM sodium orthovanadate, 1 μg/ml microcystin-LR, protease inhibitor cocktail tablet and 10% Triton X-100) to a final concentration of 1X, and cleared homogenate was further supplemented with 0.1% (w/v) SDS and 0.5% (w/v) sodium deoxycholate. The cleared homogenate was incubated on ice for 1 h for complete lysis, and centrifuged at 17 000g for 20 min at 4°C. The supernatant from this centrifugation step was designated as the total fraction. The remainder of the cleared homogenate was centrifuged at 12 000g for 10 min at 4°C. The pellets were resuspended in 15% (v/v) sedimentation buffer supplemented with the 10X lysis buffer to a final concentration of 1X (as described above) to the volume of the original sample that was spun down. The resuspended pellets were termed crude mitochondrial fractions. The supernatant from the aforementioned centrifugation step was quantified using the Bradford method and 20 mg of supernatant was ultracentrifuged at 100 000g for 1 h at 4°C. The supernatant from this centrifugation step was designated the cytosol fraction. The pellets were washed twice with PBS and resuspended in 530 μl resuspension buffer (15% w/v sedimentation buffer supplemented with 10X lysis buffer). Resuspended pellets were sonicated three times for 15 s and designated as microsomal fractions. Forty micrograms of the total fractions and the equivalent volumes of the remaining fractions were analyzed by immunoblotting.

Cell culture, treatments and lysis
MEFs were cultured in Dulbecco's modified eagle medium (DMEM) supplemented with 20% (v/v) fetal calf serum, 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, non-essential amino acids and 1 mM sodium pyruvate. Primary lung fibroblasts were cultured in Dulbecco's modified eagle medium: nutrient mixture F-12 (DMEM/F-12) supplemented with 20% (v/v) fetal calf serum, 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, non-essential amino acids and 1 mM sodium pyruvate. Generation of A549 Rab29 knock-out cells was described previously [14], and A549 cells were cultured in DMEM supplemented with 10% (v/v) fetal calf serum, 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. All cells were grown at 37°C and 5% CO 2 in a humidified atmosphere. Cell lines utilized for this study were tested regularly for mycoplasma contamination and confirmed as negative prior to experimental analysis. Cells were lysed in ice-cold lysis buffer containing 50 mM Tris-HCl pH 7.4, 1 mM EGTA, 10 mM 2-glycerophosphate, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 270 mM sucrose, supplemented with 1 μg/ml microcystin-LR, 1 mM sodium orthovanadate, complete EDTA-free protease inhibitor cocktail (Roche) and 1% (v/v) Triton X-100. Lysates were clarified by centrifugation for 15 min at 15 000g at 4°C. Protein concentrations of cell lysates were determined using the Bradford assay. MLi-2 inhibitor was dissolved in sterile DMSO, and treatment of cells with MLi-2 was for 90 min at a final concentration of 100 nM, unless otherwise indicated. Nigericin and monensin were dissolved in absolute ethanol and chloroquine diphosphate salt was dissolved in sterile water. Cell treatments were added at a dilution of 0.1% to cells. LLOMe was prepared in DMSO at a working concentration of 100 mM and cells were treated with a 1% dilution of stimulant. An equivalent volume of diluent was added to vehicle control cells where appropriate.

Generation of MEFs
Wildtype, heterozygous and homozygous Rab29 knock-out or transgenic Rab29 overexpressing MEFs were isolated from littermate-matched mouse embryos at day E12.5, as described in a previous study [55]. The resulting embryo genotypes were produced following crosses between Rab29 knock-out/wildtype (

MLi-2 inhibitor washout in MEFs
Primary MEFs of the indicated genotypes in Figure 9 were seeded in a six-well format for treatment the following day. At 50-60% confluence, MEFs were treated with vehicle (DMSO) or 100 nM MLi-2 for 48 h. To remove the MLi-2 inhibitor, cells were washed three times with warm, complete media; 15-20 min following the initial washes, cells were washed an additional two times with complete media. Cells were harvested with complete lysis buffer 30-360 min following the initial washes.  )). Cells were washed three times with 0.2% BSA for 5 min each. The slides were rinsed in sterile water before mounting on VWR super-premium microscope slides (cat# 631-0117) using Vectashield antifade mounting media (H-1000). All incubation steps were carried out at room temperature. Images were acquired using a Carl Zeiss LSM710 laser scanning confocal microscope (Carl Zeiss) using the 63X Plan-Apochromat objective (NA 1.4) and a pinhole chosen to provide a uniform optical section thickness in all fluorescence channels. The images were processed as a batch using ImageJ and brightness and contrast adjustments were kept constant for the same fluorescent channel.

Generation of primary lung fibroblasts
Primary lung fibroblasts were derived from adult mice as previously described [56]. Briefly, lung tissue was harvested and transferred to a sterile tissue culture dish, washed twice with PBS and minced with a scalpel. Tissue fragments were transferred to a 75 cm 2 cell culture flask with suitable aeration and immersed in 10 ml DMEM/ F12 supplemented with 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin and 0.14 Wunsch units/ ml collagenase (Liberase, Sigma-Aldrich 05401119001). The tissue fragments were digested with collagenase in a shaking incubator at 37°C and 5% CO 2 for 1 h. Twenty milliliters of DMEM/F12 supplemented with 20% (v/v) fetal calf serum, 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin was added, and the cell/ tissue suspension was centrifuged at room temperature at 525g for 5 min. The pellet was gently resuspended in fresh complete media and plated in a collagen-coated cell culture flask. To coat the cell culture flasks, Gibco Collagen I, rat tail (A10483-01) was diluted to a final concentration of 50 μg/ml in sterile-filtered 20 mM acetic acid. Tissue culture flasks were coated with diluted collagen according to the manufacturer's. Fibroblasts required ∼5 days to emerge from tissue fragments and regular media changes. Approximately 1.5 weeks following isolation, lung fibroblasts were passaged once, then seeded in six-well plates for experimental analysis.

Statistics
Graphs were made using Graphpad Prism 7 and 8 software. Error bars indicate SD. , and the difference between means of the four biological replicates for each genotype revealed that >99% power was achieved to reject the null hypothesis at α = 0.05 for Rab12 phosphorylation in Figure 6B-D and >90% power in Figure 6E. Similar analysis for Rab10 phosphorylation revealed that >10% power in Figure 6B, >70% power in Figure 6C, >90% power in Figure 6D and >99% power in Figure 6E was achieved in order to reject the null hypothesis at α = 0.05. For the study outlined in Figure 7B-E, post hoc power calculations derived from the standard deviations of wildtype and pathogenic mutant VPS35[D620N], and the difference between means of the four biological replicates for each genotype revealed that >99% power was achieved to reject the null hypothesis at α = 0.05 for Rab12 phosphorylation in Figure 7B-E. A similar analysis for Rab10 phosphorylation revealed that >40% power in Figure 7B and >99% power in Figure 7C-E was achieved in order to reject the null hypothesis at α = 0.05.

Expression of Rab29 in mouse tissues and cells
We raised two novel Rab29 monoclonal antibodies termed MJF-30-Clone-124 and MJF-30-Clone-104 that detect endogenous Rab29 in wildtype but not in Rab29 knock-out human A549 cells ( Figure 1A). MJF-30-Clone-124 detected both mouse and human Rab29 whilst MJF-30-Clone-104 was human specific ( Figure 1A). Immunoblotting of six mouse tissues (brain, spleen, lung, kidney, large intestine and spinal cord) and three primary mouse cell lines (MEFs, lung fibroblasts and bone marrow-derived macrophages) that all express endogenous LRRK2, revealed that Rab29 is expressed ubiquitously but levels vary significantly between tissues and cells ( Figure 1B). The highest expression is observed in macrophages and spleen, whilst low expression is seen in brain and spinal cord, and intermediate expression observed in other tissues and cells. LRRK2 was proteolyzed into at least two bands in many extracts but the expression of the two major upper bands was A B Figure 1. Detection of endogenous Rab29 in human and mouse extracts with novel Rab29 monoclonal antibodies.
(A) Thirty micrograms of the indicated cell and tissue extracts were subjected to quantitative immunoblot analysis with the indicated antibodies, diluted to 1 mg/ml. Membranes were developed using the LI-COR Odyssey CLx Western Blot imaging system. (B) Various cell and tissue extracts were analyzed for Rab29 and LRRK2 expression (upper panels). Total protein levels were visualized using the Bio-Rad Oriole Fluorescent Gel Stain and exposed using the Bio-Rad Chemidoc MP Imaging System (lower panel). Total protein levels were quantified using Bio-Rad Image Lab software. Total Rab29 or total LRRK2 was quantified using the Image Studio software. Quantified data are presented as ratios of Rab29 or LRRK2 expression divided by total protein levels, and values were normalized to the average of the respective protein expression observed in MEFs.
Quantifications are representative of three independent experiments and shown as mean ± SD. also lower in brain and spinal cord, and significantly higher in other tissues and cells. This is consistent with substantial evidence that LRRK2 and Rab29 are co-expressed in the same cells (MEFs, A549 cells, macrophages, neutrophils) [14,15,43,57,58] (http://www.immprot.org/) as well as tissues (https://www.proteinatlas.org/).

Rab29 knock-out does not impact basal LRRK2-mediated phosphorylation of Rab10 and Rab12 in mouse tissues and MEFs
To investigate the role of endogenous Rab29 in regulating LRRK2 pathway activity, we obtained Rab29 knockout mice made available through the Wellcome Trust Sanger Institute, Infrafrontier EMMA mouse repository and the Michael J. Fox Foundation (see Material and methods). These mice have also been deployed in other studies mentioned above [43,44,46]. As a readout for LRRK2 pathway activity, we measured phosphorylation of Rab10 at Thr73 [7], employing a well-characterized phospho-specific antibody [51]. Consistent with previous reports [44], the Rab29 knock-out mice were viable and displayed no overt phenotypes. We analyzed wildtype, heterozygous and Rab29 knock-out MEFs (Figure 2A) as well as wildtype and Rab29 knock-out lung ( Figure 2B), spleen ( Figure 2C), kidney ( Figure 2D) and various brain sections (Supplementary Figure S1) derived from littermate 6-month-old mice treated ± MLi-2 LRRK2 inhibitor (30 mg/kg, 2 h). Results from numerous independent experiments demonstrated that there was no significant difference in levels of LRRK2-phosphorylated Rab10, quantitated as a ratio with total Rab10, in wildtype or Rab29 knock-out MEFs, lung, spleen or kidney. In MEFs, we also analyzed LRRK2-mediated phosphorylation of Rab12 at Ser105 using a previously characterized phospho-specific antibody [21], which was also shown to be unaffected by Rab29 knock-out ( Figure 2A). Rab29 knock-out also had no impact on LRRK2 expression or phosphorylation of LRRK2 at Ser935 (Figure 2, Supplementary Figure S1). As expected, MLi-2 treatment markedly reduced Rab10 phosphorylation levels in both wildtype and Rab29 knock-out mice and was accompanied by a decrease in LRRK2 Ser935 phosphorylation. We found that although MLi-2 administration reduced Ser935 phosphorylation in brain, the low basal levels of pRab10 observed were not further decreased (Supplementary Figure S1), consistent with previous findings [51,59].

Generation and characterization of Rab29 overexpressing transgenic mice
As the human genetic data point toward variants within the PARK16 locus enhancing expression of Rab29 [15,38,39], we generated a transgenic mouse strain in which Rab29 is constitutively overexpressed in all tissues. The mouse Rab29 cDNA with no epitope tags was knocked-into the ROSA26 locus, which is frequently employed to constitutively express proteins in mouse tissues [60] ( Figure 3A). The heterozygous and homozygous Rab29 transgenic mice displayed no overt phenotype at 6 months of age, which were the oldest animals studied. Rab29 protein levels in brain, spleen, lung, kidney, large intestine and spinal cord derived from wildtype and homozygous transgenic Rab29 mice were analyzed at 3.5 and 6 months of age. This revealed that the highest overexpression of Rab29 was observed in brain (7 to 9-fold), kidney (5 to 7-fold) and large intestine (23 to 25-fold), with lower levels of overexpression observed in lung (<1.5-fold), spleen (<2-fold) and spinal cord (4-fold) ( Figure 3B). The levels of other Rab proteins measured by immunoblotting (Rab8A, Rab10, Rab12, Rab32 and Rab35) were not impacted by overexpression of Rab29 ( Figure 3B). There was no marked difference in the relative levels of Rab29 expression between 3.5 and 6-month-old mice ( Figure 3B). We also analyzed Rab29 mRNA levels in wildtype, heterozygous and homozygous brain, lung and kidney from 3.5-month-old mice ( Figure 3C). Consistent with protein levels, brain displayed the highest increase in Rab29 mRNA levels, namely 2-fold in heterozygous and ∼4-fold in homozygous Rab29 transgenic mice ( Figure 3C). In transgenic homozygous kidney and lung, Rab29 mRNA levels were increased ∼3 and 2-fold, respectively ( Figure 3C). To demonstrate that the overexpressed Rab29 was functional, we undertook immunofluorescence studies of Rab29 knock-out and littermate-matched wildtype and transgenic MEFs ( Figure 4A). This revealed that the overexpressed transgenic Rab29 was correctly localized to the Golgi, similar to wildtype Rab29, and this was confirmed by co-staining with the ACBD3 Golgi marker. Consistent with our localization being specific, no Rab29 signal was observed in the Rab29 knock-out cells and enhanced levels of Rab29 were clearly observed in the transgenic cells ( Figure 4A). Rab29 localization to the Golgi complex indicates that the protein is properly folded and active [30]. Rab29 binds to the N-terminus of LRRK2 and recent work suggests that the binding site is located within the first 552 residues of LRRK2 [41]. To obtain further evidence that the Rab29 expressed in the transgenic mouse was functional, we undertook affinity purification studies using a fragment of LRRK2 encompassing residues 1-552 in intestinal extracts that express the highest levels of Rab29. These studies confirmed that the transgenic Rab29 expressed in these extracts interacts with the LRRK2 N-terminal fragment, providing further evidence that the transgenic Rab29 is functionally competent ( Figure 4B). Binding of endogenously expressed wildtype Rab29 to the LRRK2 N-terminal fragment was also observed in wildtype intestine extracts ( Figure 4B).

Transgenic overexpression of Rab29 does not impact basal LRRK2-mediated phosphorylation of Rab10 and Rab12 in mouse tissues and cells
We examined LRRK2-mediated Rab10 and Rab12 phosphorylation in wildtype and homozygous transgenic Rab29 mouse brain ( Figure 5A), large intestine ( Figure 5B), kidney ( Figure 5C), spinal cord ( Figure 5D), spleen ( Figure 5E) and lung ( Figure 5F) from 6-month-old mice treated for 2 h ± MLi-2 LRRK2 inhibitor (30 mg/kg). As expected, phosphorylation of Rab10 and Rab12 were robustly detected in all tissues and were significantly lowered by MLi-2 treatment ( Figure 5). Notably, in the brain extracts, Rab12 phosphorylation was more sensitive to MLi-2 administration than Rab10 phosphorylation ( Figure 5A). No significant increases in transgenic Rab29 overexpressing mice were processed for total RNA extraction, cDNA synthesis, and subsequent quantitative real-time RT-PCR analysis. Rab29 mRNA levels were normalized to β-actin mRNA levels and further normalized to the wildtype of the appropriate tissue. Quantified data are presented as mean ± SD of four biological replicates per genotype. LRRK2-mediated phosphorylation of Rab10 or Rab12 were observed in any of the tissues of the Rab29 transgenic mice. Levels of LRRK2 or phosphorylation at Ser935 were also not impacted by Rab29 overexpression. Similar results were also observed in 3.5-month-old animals (Supplementary Figure S2). We also studied MEFs and lung fibroblasts derived from wildtype and homozygous Rab29 transgenic mice in which Rab29 levels were increased ∼4-fold in both cell types, and here again observed no significant impact on LRRK2-mediated Rab10 or Rab12 phosphorylation (Supplementary Figure S3). As expected, MLi-2 reduced Rab10 and Rab12 phosphorylation (Supplementary Figure S3).

Knock-out of Rab29 does not reduce elevated Rab10 and Rab12 phosphorylation in LRRK2[R1441C] knock-in mice
To study whether endogenous Rab29 was necessary for the previously reported elevated Rab10 phosphorylation observed in LRRK2[R1441C] knock-in MEFs and mouse tissues [26,51], we generated LRRK2[R1441C] knock-in MEFs as well as 6-month-old LRRK2[R1441C] ± Rab29 knock-out mice. As a control, we also generated matched LRRK2 wildtype ± Rab29 knock-out MEFs for comparison. In MEFs, as reported previously [51], the LRRK2[R1441C] knock-in mutation enhanced Rab10 phosphorylation around 3-fold compared with wildtype ( Figure 6A). Knock-out of Rab29 had no impact on Rab10 phosphorylation ( Figure 6A). As reported previously [61], the R1441C mutation reduced Ser935 phosphorylation, which was also not impacted by Rab29 knock-out ( Figure 6A). In mouse brain ( Figure 6B), lung ( Figure 6C), kidney ( Figure 6D), spleen ( Figure 6E), large intestine (Supplementary Figure S4A) and spinal cord (Supplementary Figure S4B), we also observed that knock-out of Rab29 in LRRK2[R1441C] knock-in mice had no significant impact on Rab10 phosphorylation. The LRRK2[R1441C] mutation enhanced Rab10 phosphorylation between 1.5 and 2-fold in the lung ( Figure 6B), kidney ( Figure 6C), spleen ( Figure 6D) and large intestine (Supplementary Figure S4A). No significant increase in Rab10 phosphorylation was observed in the brain ( Figure 6B) and spinal cord (Supplementary Figure S4B) of LRRK2[R1441C] knock-in mice. However, moderate sensitivity to MLi-2 was observed in the LRRK2[R1441C] knock-in mice, which was less apparent in wildtype animals ( Figure 6B). Interestingly, pRab12 in the brain samples was more clearly regulated by LRRK2[R1441C] compared with Rab10 ( Figure 6B). MLi-2 markedly decreased pRab12 levels 2.6-fold in the LRRK2[R1441C] and more modestly in wildtype brain samples, around 1.3-fold ( Figure 6B). We observed ∼2-fold increase in pRab12 levels in LRRK2 [R1441C] brain samples compared with wildtype ( Figure 6B). In contrast, phosphorylation of Rab10 was not significantly impacted by the LRRK2[R1441C] mutation or by MLi-2 in the same extracts ( Figure 6B). Similar results were also observed in the lung ( Figure 6C), kidney ( Figure 6D) and spleen ( Figure 6E). Moreover, pRab10 or pRab12 levels were not significantly impacted by knock-out of Rab29 in these tissues. We also immunoblotted for the pRab specific phosphatase PPM1H [33], and found that this was most highly expressed in the brain tissue, but found that its levels were not impacted by loss of Rab29 in multiple tissues ( Figure 6B-E). In addition, we immunoblotted kidney tissue for the M6PR that had previously been reported to be more highly expressed in Rab29 knock-out mice [46]. We confirm this finding and found that M6PR was expressed at ∼2.5 fold higher levels in the Rab29 knock-out kidney, but the expression was not significantly affected by the LRRK2[R1441C] mutation or MLi-2 administration ( Figure 6D). M6PR is often overexpressed when its trafficking is compromised.
We also attempted to measure endogenous phosphorylation of LRRK2 at Ser1292, an autophosphorylation site, using a commercially available antibody [10,54]. This site is challenging to detect, as the stoichiometry of phosphorylation at this site is believed to be low. In addition to attempting to measure endogenous LRRK2 Ser1292 phosphorylation in total extracts, we also purified microsomal enriched fractions from lungs (that express the highest levels of LRRK2), a method that was reported to facilitate detection of endogenous Ser1292 [54]. We were unable to robustly detect and quantify LRRK2 Ser1292 phosphorylation in either the total or microsome enriched fractions under conditions in which pRab10 levels were strongly detected (Supplementary Figure S5).

Knock-out of Rab29 does not reduce elevated Rab10 and Rab12 phosphorylation in VPS35[D620N] knock-in mice
To investigate whether endogenous Rab29 was necessary for the elevated Rab10 phosphorylation observed in VPS35[D620N] MEFs and mouse tissues [21], we generated VPS35[D620N] knock-in MEFs and 6-month-old VPS35[D620N] ± Rab29 knock-out mice. As a control, we also generated matched VPS35 wildtype ± Rab29 knock-out animals. As reported previously [21], VPS35[D620N] knock-in mutation enhanced Rab10 and Rab12 phosphorylation to a greater extent than is observed with the LRRK2[R1441C] pathogenic mutation (compare Figure 6 with  Figure S6B). In brain extracts derived from VPS35[D620N] mice, a moderate 1.5-fold increase in Rab10 phosphorylation was observed, which decreased with MLi-2 administration, but was also not significantly impacted by Rab29 knock-out ( Figure 7B). In contrast, the VPS35[D620N] mutation increased Rab12 phosphorylation ∼3-fold in brain extracts, which was suppressed by MLi-2 administration ( Figure 7B). The VPS35[D620N] mutation also enhanced Rab10 and Rab12 phosphorylation in the lung ( Figure 7C), kidney ( Figure 7D) and spleen ( Figure 7E). Here again, knock-out of Rab29 had no impact on Rab10 or Rab12 phosphorylation in any tissue studied. Consistent with previous results [21], the phosphorylation of Ser935 was not impacted by the VPS35[D620N] mutation. Knock-out of Rab29 had no impact on Ser935 phosphorylation in these MEFs and mouse tissues (Figure 7). We also found that M6PR was expressed at 2.5-fold higher levels in the Rab29 knock-out kidney and are increased ∼1.7-fold in the VPS35[D620N] mice ( Figure 7D).
Monovalent cation ionophore antibiotics nigericin and monensin stimulate LRRK2-mediated phosphorylation of Rab10 and Rab12 in wildtype and Rab29 knock-out cells We next investigated whether we could identify agonists that stimulate LRRK2 pathway activity. We profiled a panel of agonists and stressors, which led to the finding that structurally related antibiotics nigericin (2 μM) and monensin (10 μM) markedly enhanced phosphorylation of Rab10 and Rab12 in wildtype MEFs ( Figure 8A,B). This was blocked by treatment with the inhibitor MLi-2, indicating that nigericin and monensin stimulated Rab10 and Rab12 phosphorylation via LRRK2. Nigericin is derived from Streptomyces hygroscopicus [62], and monensin from Streptomyces cinnamonensis [63]. These agents function as monovalent cation ionophores inducing pleiotropic effects on vesicle trafficking responses [63]. We found that treatment of MEFs with 2 μM nigericin, over a 2-8 h time course, enhanced Rab10 phosphorylation 3.5 to ∼6-fold and Rab12 phosphorylation ∼6 to 9-fold ( Figure 8A). 10 mM monensin over this period enhanced Rab10 phosphorylation up to ∼2.3-fold and Rab12 phosphorylation up to ∼3.5-fold in MEFs ( Figure 8B). The knock-out of Rab29 had no significant impact on the stimulation of Rab10 and Rab12 phosphorylation observed with nigericin or monensin ( Figure 8A,B). We also found that nigericin and monensin significantly stimulated Rab10 phosphorylation in A549 cells in a manner that was also unaffected by CRISPR knock-out of Rab29 (Supplementary Figure S7). Recent work has reported that lysosomotropic agents including chloroquine [64] and the peptide LLOMe (L-leucyl-L-leucine methyl ester) [65] enhance Rab10 phosphorylation [57,[66][67][68]. We confirmed that both 50 mM chloroquine and 1 mM LLOMe (concentrations used in previous studies) enhance Rab10 as well as Rab12 phosphorylation in manner that was suppressed by MLi-2 ( Figure 8C,D). The enhancement of Rab10 and Rab12 phosphorylation by 50 μM chloroquine over an 8-h time period was up to ∼3.5 and 4-fold, respectively ( Figure 8C). Increasing the chloroquine concentration to 100 μM resulted in a 4 to 4.5-fold increase in pRab10 and pRab12 over 8 h (Supplementary Figure S8D). The increase in Rab10 and Rab12 phosphorylation upon 1 mM LLOMe stimulation approached ∼2-fold in MEFs ( Figure 8D). The knock-out of Rab29 had no significant effect on Rab10 or Rab12 phosphorylation induced by chloroquine or LLOMe ( Figure 8C,D). For the 1 mM LLOMe stimulation, we limited our analysis to 2 h, because MEF cells started to detach from tissue culture plates by 4 h. Similar results for Nigericin, Monensin, Chloroquine and LLOMe were obtained in an independent experiment with each condition undertaken in duplicate (Supplementary Figure S8). Statistical analysis of data from both sets of experiments (n = 4 datapoints for each condition-MLi-2) are included in Supplementary Figure S8.
Rate of recovery of Rab10 phosphorylation after washout of MLi-2 LRRK2 inhibitor is not impacted by Rab29 knock-out We next investigated whether endogenous Rab29 affected the rate at which Rab10 was re-phosphorylated following washout of MLi-2 in MEFs. We treated wildtype ( Figure 9A Figure 9C) ± Rab29 knock-out MEFs with MLi-2 to reduce pRab10 to undetectable levels (100 nM, 48 h). MLi-2 was removed by sequential changes of medium over a 15 min period, and phosphorylation of Rab10 quantified at time points up to 6 h. Under these conditions, we observed that the rate of recovery in the three cell lines was not affected by Rab29 knock-out. For the wildtype and VPS35[D620N] MEFs we observed 60-70% recovery of Rab10 phosphorylation within 6 h. However, in the LRRK2[R1441C] MEFs, recovery of pRab10 was only ∼30% after 6 h, and it is possible that the higher affinity of MLi-2 for this pathogenic mutant might account for this.

Evidence that Rab32 is not compensating for loss of Rab29 in regulating LRRK2
Finally, we explored whether Rab32, which is closely related to Rab29 and reported to interact with LRRK2 [41,69], could contribute to the regulation of LRRK2 activity in Rab29 knock-out MEFs. siRNA knockdown reduced Rab32 expression by 70-80%, however, this had no impact on Rab10 phosphorylation in Rab29 knockout wildtype LRRK2 or LRRK2[R1441C] knock-in cells ( Figure 10). Knockdown of LRRK2 in parallel experiments, as expected, markedly reduced Rab10 phosphorylation ( Figure 10). By immunoblotting analysis of wildtype and Rab29 knock-out MEFs we were unable to detect Rab38, which is also related to Rab29 and Rab32. This is consistent with previous high-resolution proteomic analysis of MEFs in which Rab38 was not detected [7].

Discussion
As outlined in the introduction, considerable evidence supports the view that transient overexpression of Rab29 recruits the bulk of cellular LRRK2 to the Golgi surface, leading to its activation. In overexpression studies, the LRRK2[R1441C/G] mutants are more readily activated by Rab29, which could explain why this mutation elevates LRRK2 activity. However, our results suggest that knock-out of endogenous Rab29 has no significant inhibitor was removed from cells through multiple washes with complete media for the indicated periods prior to cell lysis.
Twenty micrograms of whole-cell extracts were subjected to quantitative immunoblot analysis with the indicated antibodies.
The membranes were developed using the LI-COR Odyssey CLx Western Blot imaging system. Quantified data are presented as the ratios of phospho-Rab10/total Rab10, which were calculated using the Image Studio software. Values were normalized to the average of (A) wildtype MEFs treated with DMSO, (B) LRRK2[R1441C] MEFs treated with DMSO, or (C) VPS35[D620N] MEFs treated with DMSO. Cell extract derived from a wildtype MEF cell line (**) was added to each gel in order to accurately compare the Rab10 pThr73/Rab10 total ratios between genotypes in each panel. Data quantifications are presented as mean ± SD. Similar results were obtained in two separate experiments.
impact on endogenous LRRK2 activity, assessed by monitoring pRab10 and pRab12 levels in six mouse tissues as well as MEFs and lung derived fibroblasts. Moreover, we also found that knock-out of Rab29 does not impact elevated Rab10 or Rab12 phosphorylation observed in the LRRK2[R1441C] knock-in MEFs or mouse tissues. This would suggest that endogenous Rab29 is not sufficient to explain the elevated activity of the LRRK2[R1441C] pathogenic mutant. We also found that Rab29 knock-out had no effect on the elevated LRRK2-mediated Rab10 or Rab12 phosphorylation observed in VPS35[D620N] knock-in MEFs and mouse tissues. To our knowledge, there is no evidence implicating Rab29 in mediating the effects of the VPS35 [D620N] mutation. We also find that in brain extracts, Rab12 phosphorylation appears to be more robustly impacted by LRRK2 inhibitors and pathogenic mutations than Rab10 phosphorylation, consistent with a recent study [70].
In this study, we also report that agents that induce endolysosomal pathway stress including nigericin, monensin, chloroquine and LLOMe significantly enhance Rab10 and Rab12 phosphorylation in MEFs and A549 cells, in a manner that is blocked by LRRK2 inhibitors. Out of the 4 agents tested, nigericin was the most The indicated matched primary MEFs were transfected with Dharmacon smartPOOL siRNA targeting either LRRK2, Rab32, or control scrambled non-targeting siRNA. The cells were lysed 72 h post-transfection. 20 mg of whole-cell extracts were subjected to quantitative immunoblot analysis with the indicated antibodies. The membranes were developed using the LI-COR Odyssey CLx Western Blot imaging system. Quantified data are presented as the ratios of phospho-Rab10/Total Rab10 and were calculated using the Image Studio software. Values were normalized to the average of Rab29 knock-out MEFs treated with scrambled siRNA. The ratio of LRRK2 or Rab32 expression divided by the loading control was used to determine siRNA knockdown efficiency, and values were normalized to the average of the scrambled siRNA treated MEFs. Data quantifications are represented as mean ± SD. Similar results were obtained in three separate experiments. potent in enhancing Rab10 and Rab12 phosphorylation and increases the basal phosphorylation of these substrates 5 to 9-fold (Figure 8, Supplementary Figure S8). Nigericin and monensin function as broad monovalent cation ionophores, and have been reported to have pleiotropic effects on vesicular trafficking pathways [63]. Recent reports have also shown that chloroquine and LLOMe enhance Rab10 phosphorylation [57,[66][67][68], and we find that these agents also promote Rab12 phosphorylation ( Figure 8C,D, Supplementary Figure S8). We find that in MEFs, monensin, chloroquine and LLOMe induce pRab10 and pRab12 around 2 to 4-fold, compared with 5 to 9-fold effects observed with nigericin. A recent report suggested that chloroquine induced phosphorylation of Rab10 in RAW264.7 macrophages in a manner that was largely blocked by siRNA knockdown of Rab29 [67]. In contrast, we find that in MEFs, knock-out of Rab29 does not impact moderate Rab10 or Rab12 phosphorylation induced by chloroquine as well as the other agents we have tested (nigericin, monensin, LLOMe). It should be noted that Rab29 is highly expressed in macrophages [47] and in future work, it would be important to utilize Rab29 knock-out models to verify whether Rab29 plays a role in regulating LRRK2 pathway activity through agents such as chloroquine in primary macrophages. Further work is required to understand the mechanism by which nigericin, monensin and lysosomotropic agents promote LRRK2-mediated Rab protein phosphorylation. It would be interesting to explore whether these agonists promote the recruitment of LRRK2 to particular membrane compartments, thereby activating LRRK2 and promoting Rab10 and Rab12 phosphorylation at that location. Recruitment of LRRK2 to membranes has been proposed previously to be a key mechanism by which LRRK2 activity is regulated [29,57,68].
It is possible that other Rab proteins, or even other regulators that have not yet been characterized, could also interact with and activate LRRK2 in a similar manner to Rab29 by recruiting LRRK2 to a variety of cellular membranes. In the absence of Rab29 overexpression, LRRK2 is widely distributed in cells with ∼90% cytosolic localization and ∼10% localization on a variety of cellular membranes [14,57,71,72]. A small fraction of LRRK2 appears to be localized in the Golgi region without Rab29 overexpression [14][15][16][17]73], which may explain why Rab29 knock-out does not have a noticeable impact on Rab10 or Rab12 protein phosphorylation when measured in a whole cell or tissue extract. In future work, it will be important to develop assays in which the pool of endogenous LRRK2 that resides at the Golgi could be specifically assessed. Since LRRK2 is likely in equilibrium between membranes and the cytosol [29], such interactions may be transient and more challenging to capture quantitatively. Indeed, by elevating the local concentration of Rab29 on the Golgi by exogenous expression, this pool was more readily detected. It will be important to define whether the activity and localization of the endogenous pool of Golgi-resident LRRK2 is dependent upon Rab29. It will also be necessary to further study whether endogenous LRRK2 located on other specific membrane compartments relies on other Rab proteins or regulators for this localization, and whether this contributes to the total cellular LRRK2 activity measured in cell extracts. It would also be important to define more precisely the residues in LRRK2 that bind Rab29 and investigate how subtle mutations that prevent Rab29 binding, impact cellular LRRK2 activity. Whether other Rab proteins bind to the same or different sites in LRRK2 should also be investigated.
The ability of exogenous Rab29 to recruit LRRK2 to the Golgi (or other compartments to which it is targeted) confirms the ability of Rab29 to bind LRRK2 in cells. Yet our knock-in and knock-out models failed to reveal clues regarding the functional significance of this interaction. Rab29 is a relatively poorly abundant Rab and may play a specific role in macrophages and dendritic cells where it is most abundant. Indeed, LRRK2 activation triggered by Rab29 could occur in a specific cell type or tissue, or following a physiological stimulus, stress or infection that we have not investigated. Our data do not rule out the possibility that Rab29 knock-out is compensated by another cellular protein(s) other than Rab32. However, our findings with transgenic mice that overexpress Rab29 from 1.5 to 25-fold, without enhancing LRRK2-mediated Rab10 or Rab12 phosphorylation, reveal that increasing Rab29 expression is not sufficient to stimulate the activity of endogenous LRRK2.
We have previously shown that Rab3, Rab8A, Rab10 Rab12, Rab35 and Rab43 are the main substrates of LRRK2 [7,26,74], and have herein monitored changes in phosphorylation of Rab10 and Rab12. We, therefore, cannot exclude the possibility that Rab29 preferentially activates phosphorylation of the Rab proteins we have not assayed, but we consider this unlikely because this set of Rab proteins appear to be co-ordinately phosphorylated in all cultured cell experiments that we have carried out to date [26,33,51]. It has also been suggested that screens could be undertaken to identify potentially therapeutic compounds that block Rab29 binding to LRRK2, however, our data suggest that such agents may not be effective at reducing basal LRRK2 activity, unless these chemicals also block LRRK2 binding to other regulators that transport it to cellular membranes. Finally, we propose that the new Rab29 monoclonal antibodies we have developed could be exploited to better understand how Parkinson's mutations within the PARK16 locus impact Rab29 protein expression.