Autosomal-dominant, missense mutations in the leucine-rich repeat protein kinase 2 (LRRK2) gene are the most common genetic predisposition to develop Parkinson's disease (PD). LRRK2 kinase activity is increased in several pathogenic mutations (N1437H, R1441C/G/H, Y1699C, G2019S), implicating hyperphosphorylation of a substrate in the pathogenesis of the disease. Identification of the downstream targets of LRRK2 is a crucial endeavor in the field to understand LRRK2 pathway dysfunction in the disease. We have identified the signaling adapter protein p62/SQSTM1 as a novel endogenous interacting partner and a substrate of LRRK2. Using mass spectrometry and phospho-specific antibodies, we found that LRRK2 phosphorylates p62 on Thr138 in vitro and in cells. We found that the pathogenic LRRK2 PD-associated mutations (N1437H, R1441C/G/H, Y1699C, G2019S) increase phosphorylation of p62 similar to previously reported substrate Rab proteins. Notably, we found that the pathogenic I2020T mutation and the risk factor mutation G2385R displayed decreased phosphorylation of p62. p62 phosphorylation by LRRK2 is blocked by treatment with selective LRRK2 inhibitors in cells. We also found that the amino-terminus of LRRK2 is crucial for optimal phosphorylation of Rab7L1 and p62 in cells. LRRK2 phosphorylation of Thr138 is dependent on a p62 functional ubiquitin-binding domain at its carboxy-terminus. Co-expression of p62 with LRRK2 G2019S increases the neurotoxicity of this mutation in a manner dependent on Thr138. p62 is an additional novel substrate of LRRK2 that regulates its toxic biology, reveals novel signaling nodes and can be used as a pharmacodynamic marker for LRRK2 kinase activity.
Parkinson's disease (PD) is a progressive neurodegenerative disorder with no known cure. PD is typically of idiopathic origin; however, it has been established that environmental exposures to toxins and inheritance of dominant or recessive mutations can precipitate the onset of disease. The neuropathological hallmark of PD is the presence of cytoplasmic α-synuclein inclusions [1,2]. Cytoplasmic proteinaceous toxic aggregates are common features of multiple neurodegenerative disorders. Selective autophagy directs the clearance of aggregated proteins and dysfunctional organelles, and it is thought that this process is necessary for handling these toxic aggregates. It is likely that autophagic handling of dysfunctional or aggregated proteins is disrupted in neurodegenerative diseases, and could be due to deregulation of protein chaperones or autophagy adapter proteins.
The adapter protein p62/sequestosome-1 (p62) is a component of cytoplasmic ubiquitin-positive inclusions in PD and several protein aggregation-based neurological disorders [3–11]. The p62 protein has multiple characterized domains [12–14] including a PB1 domain (N′-Phos and Bem1 domain) that enables self-oligomerization and binds signaling molecules, an LC3 interaction region (LIR) and a ubiquitin-binding domain (UBA) through which p62 binds ubiquitinated protein aggregates for sequestration and formation of the autophagosome. Other protein interaction domains are a TRAF6-binding domain (TB), nuclear import and export signals (NLS and NES), a KEAP interaction region (KIR) and a zinc finger (ZZ) domain. p62 itself is an autophagy substrate, and inhibition of the autophagy machinery causes accumulation and oligomerization of p62 [15–17]. The several functions and pathologies of p62 function are regulated by post-translational modifications such as ubiquitination and phosphorylation [15,18]. p62 activity bridges autophagy to the stress response pathway through regulation of the interaction of KEAP with Nrf2 to promote its stability and gene activation. This interaction is regulated by p62 phosphorylation at Ser349 by mammalian target of rapamycin complex 1 (mTORC1), casein kinase 1 (CK1) and TAK1 [19–21]. p62 facilitates the removal of intracellular organelles dependent on activation and phosphorylation by TBK1 on Ser403 [22,23]. CK2 also regulates Ser403, increasing affinity for ubiquitin to regulate the selective clearance of ubiquitinated proteins . ULK1 phosphorylation of Ser407 regulates phosphorylation of Ser403 in an interplay of autophagy kinases . Additionally, p62 is phosphorylated by CDK1 at Thr269 , and p38δ phosphorylation of p62 Thr269 is necessary for amino acid-dependent activation of mTORC1 . PB1 domain interaction partner selection can be regulated by phosphorylation by PKA on Ser24 .
Autosomal-dominant, missense mutations in the leucine-rich repeat protein kinase 2 (LRRK2) gene are a genetic predisposition to develop PD [29–33]. LRRK2 mutations account for ∼1–5% of familial and sporadic PD and are inherited with an autosomal-dominant pattern with incomplete penetrance [34–39]. The most common mutation leads to a serine substitution of Gly2019 in subdomain VII of the kinase domain , which increases kinase activity 2- to 4-fold [40–42]. Other pathogenic inherited mutations in the Roc/COR domain (R1441G/C/H, Y1699C and N1437H) also result in increased kinase activity, but are dephosphorylated at Ser910/935/955/973 . It is currently unknown how pathogenic mutations in LRRK2 across its multiple domains cause PD. Inhibition of LRRK2 kinase activity and several pathogenic LRRK2 mutations have been shown to alter LRRK2 protein complexes, with a dynamic loss and gain of binding partners, and redistribution within the cell.
Interestingly, cytoplasmic aggregates of LRRK2 expressed in cell culture and primary neurons co-localize with p62 [44,45]. Furthermore, overexpression of LRRK2 induces the accumulation of p62 [46,47], while LRRK2 knockouts similarly increase p62 accumulation [48,49], indicating a potential signaling axis for LRRK2 in regulating autophagy. LRRK2 has been associated with dysfunctions in multiple cellular processes , for example translation [51,52], mitochondrial health [53–56], vesicular trafficking and autophagy [45,57–66], cytoskeletal organization [67–72], WNT signaling [73,74], NFAT signaling  and inflammation. Endogenous LRRK2 phosphorylation of a subset of Rabs implicates LRRK2 function in vesicular trafficking and autophagy . Several studies link LRRK2 function to autophagy; however, the mechanism of how enhanced kinase activity, through mutation, alters protein degradation pathways has yet to be validated. However, LRRK2 kinase inhibition stimulates autophagy in several systems, implicating substrate modification in the process [76–78]. Interestingly, it was reported that activated Nrf2 relieves some of the neuronal toxicity presented by overexpression of mutant LRRK2 , and p62 links autophagy and the Nrf pathway [19,80]. In an attempt to identify a novel mechanism of LRRK2 function in stress response and autophagy, we investigated the functional interaction of LRRK2 with p62. We confirmed the interaction of LRRK2 and p62, and mapped the reciprocal sites of interaction in cells and in vitro. Proteins in complex with kinases are also sometimes substrates for these enzymes, and we found that p62 is phosphorylated by LRRK2 in a manner that depends on the LRRK2 amino-terminus and phosphorylation on Ser910/935. LRRK2 (G2019S) neuronal toxicity is enhanced by co-expression with p62 in a manner dependent on Thr138. These data have important implications for the role of LRRK2 in neuronal stress response and death through the regulation of p62 .
Materials and methods
Buffers, chemicals and antibodies
Lysis buffer contained 50 mM Tris–HCl (pH 7.4), 1 mM EGTA, 1 mM EDTA, 1 mM sodium orthovanadate, 10 mM sodium β-glycerolphosphate, 50 mM NaF, 5 mM sodium pyrophosphate, 0.27 M sucrose, 1 mM benzamidine and 1 mM phenylmethanesulphonylfluoride (PMSF), and was supplemented with Sigma Protease Inhibitor Cocktail (Sigma) and ROCHE PhosStop (Roche). Detergents used in the lysis buffer are 1% Triton X-100 or 1% NP40. Lysis buffer used for endogenous p62/SQSTM1 pThr138 detection was supplemented with 0.15 M NaCl and 0.5% NP40. Buffer A contained 50 mM Tris–HCl (pH 7.4), 50 mM NaCl, 0.1 mM EGTA and 0.27 M sucrose. Proximity ligation assay reagents were from Duolink. LRRK2 kinase inhibitor GNE1023 was described in ref. , and synthesized at and provided by Genentech; PF475 was described in ref.  and purchased from Sigma. Anti-GFP (clones 7.1 and 13.1) and anti-HA (clone 3F10) antibodies were bought from Roche. Beta-actin (D6A8) and Hsp90 (C45G5) were from Cell Signaling Technologies. Anti-LRRK2 (N241) is from Neuromab. Anti-LRRK2 pSer935 (UDD2) and anti-LRRK2 pSer1292 were from Abcam. Rabbit polyclonal anti-p62 phosphothreonine 138 (pThr138) antibody was generated by injection of the KLH-conjugated phosphopeptide NGPVVGpTRYKC*SV (where pT is phosphothreonine and * indicates the KLH conjugation site) into rabbits and was affinity-purified by positive and negative selection against the phospho- and de-phosphopeptides, respectively, at Yenzym Inc. Anti-pThr138 was used at a final concentration of 1 µg/ml in the presence of 10 µg/ml non-phosphorylated peptide. Mouse anti-p62 (M162-3), rabbit anti-p62 (PM045) and rat anti-phospho-p62 Ser403 (D343-3) were from MBL. FLAG M2 antibody was purchased from Sigma. Mouse anti-β3 tubulin (Tuj1) was from Biolegend (Covance). Rabbit anti-active caspase-3 was from R&D Systems. Sheep anti-phospho-Rab8, sheep anti-phospho-Rab10 and sheep anti-phospho-Rab7L1 were kind gifts from Professor Dario Alessi (MRC-PPU, University of Dundee, U.K.). Anti-GFP for immunoblotting from Roche (clones 7.1 and 13.1) or rabbit monoclonal anti-GFP from LifeTechnologies was used for proximity ligation assay (PLA). GFP-Trap agarose was purchased from Chromotek. Calyculin A and Okadaic acid were from LC Labs, and MLi-2  was a kind gift of Professor Dario Alessi (MRC-PPU, University of Dundee, U.K.).
Cell culture, treatments and cell lysis
Tissue culture reagents were from LifeTechnologies or ThermoScientific. HEK-293 cells (from ATCC) were cultured in Dulbecco's Modified Eagle's medium supplemented with 10% FBS, 2 mM glutamine and 1× antimycotic/antibiotic solution. The Flp-in T-REx system was from Invitrogen and stable cell lines were generated as per the manufacturer's instructions by selection with hygromycin as has been described previously . T-REx cell lines were cultured in DMEM supplemented with 10% FBS and 2 mM glutamine, 1× antimycotic/antibiotic and 15 µg/ml blasticidin and 100 µg/ml hygromycin. Human lung alveolar epithelial A549 cells (from ATCC) were cultured in DMEM/F-12 with l-Glutamine and 10% FBS and 1× antimycotic/antibiotic. Human control lymphoblasts were a kind gift of Dr Birgitt Schuele (The Parkinson's Institute, Sunnyvale, CA) and cultured in RPMI supplemented with l-glutamine and 10% FBS, 1× antimycotic/antibiotic and 1× non-essential amino acid (Gibco). Mouse embryonic fibroblasts (MEFs) were cultured in DMEM supplemented with l-glutamine and 10% FBS, 1× antimycotic/antibiotic and 1× non-essential amino acid (Gibco). HEK293 and T-REx were transfected by the polyethylenimine method  and were lysed 48 h after transfection. T-REx cultures were induced to express the indicated protein by inclusion of 1 µg/ml doxycycline in the culture medium for 48 h. After the indicated culture conditions, cell lysates were prepared by washing once with PBS and lysing in situ with 0.4 ml of lysis buffer per 10 cm dish on ice, and then centrifuged at 15 000×g at 4°C for 15 min. Protein concentrations were determined using the Bradford method with BSA as the standard. Generation and culture of CRISPR/Cas9 GFP-LRRK2 H1299 cells can be found in Supplementary Figure S1.
Restriction enzyme digests, DNA ligations and other recombinant DNA procedures were performed using standard protocols with Fermentas enzymes. DNA constructs used for transfection were purified from Escherichia coli DH5α using Qiagen plasmid Maxi kits according to the manufacturer's protocol. The pcDNA5-Frt-FLAG-LRRK2 and pcDNA5-Frt-GFP-LRRK2 constructs and pcDNA5-Frt-FLAG-LRRK1 used for transfections were provided by Professor Dario Alessi (MRC-PPU, University of Dundee, U.K.). The LRRK2 cDNA was subcloned from pcDNA3.1  into the pcms-EGFP reporter plasmid, expressing untagged LRRK2 and EGFP under separate promoters, or into pcDNA3.1 with a C-terminal c-myc epitope tag. A plasmid encoding HA-tagged p62 was obtained from Addgene (plasmid #28027) and was subcloned into pcDNA5-Frt-FLAG and pGEX-6P. pFLAG HDAC4 was a gift from Eric Verdin (Addgene plasmid #13821), pcDNA3 IKKε FLAG was a gift from Tom Maniatis (Addgene plasmid #26201) and HDAC6 Flag was a gift from Eric Verdin (Addgene plasmid #13823); Harm Kampinga provided pcDNA5/FRT/TO-V5 HSPA1A (Addgene plasmid #19510), pcDNA5/FRT/TO-V5 HSPH1 (Addgene plasmid #19506) pcDNA5/FRT/TO-V5 HSPA8 (Addgene plasmid #19514), pcDNA5/FRT/TO-V5 DNAJB6b (Addgene plasmid #19528), pcDNA5/FRT/TO-V5 Hsp27 (Addgene plasmid #63102) and pcDNA5/FRT/TO-V5 DNAJB8 (Addgene plasmid #19531). Hsp90 was a gift from William Sessa (Addgene plasmid #22487); pcDNA5-Frt-GFP-Rab7L1 was synthesized using codon optimization by GeneArt (ThermoScientific). DNA manipulations were carried out using standard techniques, and mutagenesis was performed with the GeneArt Mutagenesis Kit (ThermoScientific).
For transfected HEK293 or T-REx cells, cell lysates were prepared in lysis buffer (0.4 ml per 10 cm dish) and subjected to immunoprecipitation with anti-FLAG M2 agarose or GFP-Trap A beads (Chromotek) at 4°C for 2 h. Beads were washed twice with lysis buffer supplemented with 300 mM NaCl, and then twice with buffer A. Immune complexes were incubated at 70°C for 10 min in LDS sample buffer, passed through a Spin-X column (Corning) to separate the eluate from the beads and then boiled. Equal amounts of immunoprecipitated protein were subjected to western blots with indicated antibodies. For endogenous LRRK2 immunoprecipitation, mouse anti-LRRK2 (N241; Neuromab) non-covalently conjugated to protein-G agarose (1 µg antibody: 1 µl bead) was used to enrich LRRK2 protein complexes from A549 cells. Endogenous p62/SQSTM1 was immunoprecipitated by mouse anti-p62 (MBL) from human lymphoblastoid cells covalently conjugated to protein-A sepharose (1 µg antibody: 1 µl bead; Pierce Crosslink IP Kit, ThermoScientific) and incubated at 4°C for 4 h. The immunoprecipitated samples were analyzed by immunoblotting.
To induce the expression of GST-p62, E. coli BL21 Rosetta transformants of pGEX-6P plasmids harboring p62 or mutant p62 were grown to an OD600 of 0.5 at 37°C and induced at 30°C for 3 h by the addition of IPTG (isopropylβ-d-thiogalactoside) to a final concentration of 1 mM. Cells were lysed by sonication in lysis buffer with 1% (v/v) Triton X-100 and 0.1% β-mercaptoethanol. The soluble fraction was retrieved by centrifugation at 15 000 g for 15 min. Recombinant protein was purified by glutathione–Sepharose chromatography, and proteins were eluted in buffer A with 20 mM glutathione, 1 mM benzamidine and 2 mM PMSF, or were liberated from the GST fusion by incubation with 0.1 mg of precision protease per 1 mg of protein.
For assays using recombinant proteins as substrates, the reactions were set up in a total volume of 25 µl with recombinant kinase (GST-LRRK2970-2527 or FLAG-LRRK21-2527) in 50 mM Tris (pH 7.5), 0.1 mM EGTA, 10 mM MgCl2 and 0.1 mM [γ-32P] ATP (∼500 cpm/pmol), with p62 substrate, with the indicated concentrations of enzyme and substrate. After incubation for 30 min at 30°C, the reactions were stopped by the addition of Laemmli sample buffer. Reaction products were resolved by electrophoresis on NuPage Bis–Tris gels and stained with Coomassie blue. The incorporation of phosphate into protein substrates was determined by autoradiography and/or immunoblotting with phospho-specific antibodies. Kinase reactions for mass spectrometry-based phosphosite identification were carried out under similar conditions but with 200 µM cold ATP; reactions were resolved by electrophoresis on NuPage Bis–Tris gels and stained with colloidal blue, and p62 was excised and submitted for MS analysis at MS Bioworks.
Sample preparation and LC–MS/MS
Replicate gel segments were reduced using dithiothreitol, alkylated with iodoacetamide and then subjected to digestion with trypsin, chymotrypsin and elastase (Promega, Madison, WI). Digests were analyzed by nano-LC–MS/MS with a NanoAcquity HPLC system (Waters, MA) interfaced to a Q Exactive tandem mass spectrometer (ThermoFisher, San Jose, CA). Peptides were loaded on a trapping column and eluted over a 75 µm analytical column at 350 nl/min; both columns were packed with Jupiter Proteo resin (Phenomenex, Torrance, CA). A 30 min gradient was employed for each digest. The mass spectrometer was operated in a data-dependent mode, with MS and MS/MS performed in the Orbitrap at the 70 000 and 17 500 FWHM resolution, respectively. The 15 most abundant ions were selected for MS/MS from each MS scan. Dynamic exclusion and repeat settings ensured each ion were selected only once and excluded for 30 s thereafter.
Product ion data were searched against the combined forward and reverse protein database using a locally stored copy of the Mascot search engine v2.5 (Matrix Science, London, U.K.) via Mascot Daemon v2.5. Peak lists were generated using Proteome Discoverer v1.4 (ThermoFisher). The database was appended with common background proteins. Search parameters were as follows: precursor mass tolerance 10 ppm, product ion mass tolerance 0.02 Da, two missed cleavages allowed, fully tryptic peptides only for trypsin and no enzyme specificity for chymotrypsin and elastase, fixed modification of carbamidomethyl cysteine, variable modifications of oxidized methionine, protein N-terminal acetylation, pyro-glutamic acid on N-terminal glutamine and phosphorylation on serine, threonine and tyrosine.
Mascot search result flat files (DAT) were parsed to the Scaffold software v3.1 (Proteome Software, Portland, OR) to create a non-redundant list per sample. The criteria for accepting protein identification were determined by calculating the false discovery rates (FDRs) from the concatenated forward/reverse database. This resulted in the following cutoff values: 90% protein and 50% peptide level probability (probabilities were assigned by the Protein Prophet algorithm) and a minimum of two unique peptides per protein. These criteria resulted in FDRs of less than 1% at the protein level.
Immunocytochemistry and PLA
For immunocytochemistry, cells were grown on coverslips, washed with PBS and fixed with 4% paraformaldehyde for 15 min, followed by two wash steps with PBS. After permeabilization for 5 min with PBST, a 30 min blocking step with 10% goat serum (Dako cytomation) in PBS was performed. This was followed by 2 h incubation with FLAG M2 antibody in PBS. After three washes with PBS for 5 min, the cells were incubated for 1 h with secondary antibody (Alexa Fluor®-conjugated antibody, 1 : 500 dilution, Molecular Probes) and washed three times for 5 min with PBS. PLAs were performed on eight-well CC2 chamber slides. Cells were fixed and permeabilized as in the immunocytochemistry experiments, then blocked with Duolink Blocking Solution for 30 min at 37°C. Immediately after blocking, the cells were incubated with the indicated primary antibodies in the Duolink Antibody Diluent for 1 h at room temperature (20–24°C). The cells were washed twice with Duolink Wash Buffer A for 5 min before the incubation of the PLUS and MINUS PLA probes. The cells were incubated with PLA probes for 1 h at 37°C, followed by two washes with Wash Buffer A for 5 min. The cells were then incubated with premixed Ligation–Ligase solution at 37°C for 30 min. After two washes with Wash Buffer A for 2 min, the cells were incubated with premixed Amplification-Polymerase solution for 60–90 min at 37°C. Finally, the cells were washed twice with Duolink Wash Buffer B for 10 min, followed by 0.01 × Wash Buffer B for 1 min. The slides were dried in the dark and mounted with Duolink II Mounting Medium. The images were taken with a Nikon Eclipse TI florescence microscope and quantified by Duolink ImageTool (Version 126.96.36.199); ∼150 cell signals were counted for each sample. The quantification data were graphed by Prism5 (GraphPad), and the statistics were calculated by one-way ANOVAs. The quantification of the PLA results was from three independent experiments.
Primary neuronal survival assays
Primary rat embryonic cortical neurons were prepared as described recently . Briefly, embryonic day 17 rat cortices were dissociated in trypsin/DNAse followed by gentle mechanical disruption. Dissociated neurons were plated at a density of ∼125 000 cells/cm2 in complete Neurobasal medium with l-glutamine, penicillin/streptomycin and B-27 serum-free supplements (all components from ThermoScientific). On day 4 following plating, the neurons were transiently transfected with untagged human LRRK2 [wild type (WT) or G2019S] cDNA in pcms-EGFP expression vector using Lipofectamine 2000 (ThermoScientific) as described recently , together with FLAG-tagged p62 (WT or T138A) at a ratio of 3 : 1 with LRRK2 in excess. Three days following transfection, the neurons were fixed in 4% paraformaldehyde, and processed for anti-LRRK2 (MJFF c41-2) or anti-FLAG/anti-GFP immunofluorescence, with DAPI as a nuclear counterstain. Co-expression of p62 and LRRK2 within individual neurons, transfected with an excess of LRRK2 cDNA, was confirmed by co-immunostaining for anti-Flag and anti-LRRK2 (clone c41-2), and the neuronal phenotype of the cultures was established by triple immunostaining for anti-GFP, anti-Flag and anti-β3 tubulin. The stained coverslips were observed under 40× magnification, and neurons double positive for EGFP and FLAG were scored according to their nuclear morphology. Neurons were considered to be apoptotic if two or more condensed chromatin bodies were observed within a neuronal profile. From at least four parallel coverslips (performed from two independent cultures), 100 positive neurons were counted by a researcher blind to the experimental conditions, and apoptotic neurons were expressed as a percentage of neurons positive for both FLAG and EGFP. Parallel coverslips were co-immunostained for anti-Flag and anti-active caspase-3, together with DAPI, as an additional marker of neurons undergoing apoptotic cell death.
LRRK2 interacts with p62
LRRK2 is known to interact with chaperone proteins Hsp90, Cdc37, Hsp70 and Hsc70 [64,86–88]. These chaperones are necessary for proper folding of proteins. As cytoplasmic accumulations of LRRK2 have been found coincident with p62, we sought to survey several protein chaperones involved in post-translational protein handling. Using epitope tag immunoprecipitations, we screened a panel of protein chaperones for interaction with LRRK2 using plasmid-based expression in T-Rex 293-GFP or T-Rex 293-GFP-LRRK2 cells. We tested the signaling chaperone Hsp27, along with the protein-folding chaperones DNAJB6 (Hsp40), DNAJB8 (Hsp40), HSPA1A (Hsp70), HspA8 (Hsc70), Hsp90, HSPH1 (Hsp110), and also HDAC6 and p62, which are autophagy-related proteins that can shuttle cargo to aggresomes or autophagosomes. We found that LRRK2 interacts with all of these chaperones in this expression system, as well as p62 and HDAC6, except for Hsp110 (Figure 1A). To determine spatial proximity of LRRK2 and p62 in cells, we performed the in situ proximity ligation assay (PLA) on cells expressing GFP-LRRK2 and FLAG-p62 (Supplementary Figure S2A). We compared the number of signals per p62/LRRK2 co-transfected cell with the number of signals per cell transfected with cytoplasmic HDAC4 or vector alone and found a significant and specific increase in p62–LRRK2 signals (representative micrographs in Supplementary Figure S2A,B). LRRK2 and its paralog LRRK1 harbor a similar domain structure, and both could therefore interact with p62. We asked if p62 could interact with LRRK2 or LRRK1 in co-expression analyses. FLAG-tagged LRRK2 or LRRK1 was expressed with HA-p62, and anti-FLAG immunoprecipitates were analyzed for the presence of p62. LRRK1 has also been associated with autophagy . Figure 1B shows that only p62 was highly enriched in LRRK2 immune complexes, with a slight increase in the presence of LRRK2 inhibitor GNE1023.
LRRK2 interacts with p62/SQSTM1.
We next sought to isolate endogenous LRRK2/p62 complexes. To do this, we tested two cell lines and immunoprecipitation methods. We first immunoprecipitated endogenous LRRK2 immunoprecipitates from A549 cells and found endogenous p62 present (Figure 1C), where inclusion of LRRK2 inhibitor resulted in a slight increase in the levels of p62 in the immunoprecipitate. As we observed a slight increases in p62–LRRK2 complexes in cells after LRRK2 inhibition, we sought to quantitate this interaction with the selective inhibitor MLi2 and found a reproducible, but modest increase in p62 co-precipitation with inhibited LRRK2 (Supplementary Figure S2C). We next employed the PLA on A549 cells to investigate if endogenous LRRK2 and p62 were in close spatial proximity. Using specific antibodies against LRRK2 and p62, we found significant numbers of PLA signals for LRRK2/p62 complexes over IgG controls and significantly more PLA signals when treated with LRRK2 inhibitor for 90 min (Supplementary Figure S2D). To complement our endogenous interaction data with another immunological retrieval tool, we next expressed amino-terminal-tagged GFP-LRRK2 from a Bacmam virus and 24 h after infection cells were fixed at 0, 30 min and 3 h after treatment with LRRK2 inhibitor (Supplementary Figure S2E). After acute LRRK2 inhibition or in mutants that are dephosphorylated at Ser910/935 (e.g. R1441C/G/H and Y1699C), LRRK2 forms skein or punctate structures in the cytoplasm of cells. Staining for endogenous p62 revealed an increase in co-localization with LRRK2 skeins and puncta over time. Using CRISPR/Cas9, we generated an H1299 cell line targeted with GFP at the endogenous start-codon of LRRK2 to confirm the interaction of LRRK2 and p62 with an alternate tag and cell background, as described in Supplementary Figure S1. We found that endogenous GFP-LRRK2 interacts with p62. Taken together, these data show that LRRK2 and p62 interact and confirm previously reported data from Park et al.  showing endogenous interaction.
Identification of the p62 and LRRK2 interaction domains
LRRK2 and p62 are multi-domain proteins with several protein interaction regions. To provide insights into downstream biology of the interaction, we determined the interaction domain of p62 on LRRK2 and the interaction domain of LRRK2 on p62 using deletion plasmid constructs. To identify the LRRK2 interaction domain with p62, we expressed GFP-tagged LRRK2 fragments (Figure 2A) and analyzed GFP immunoprecipitates for endogenous p62 (Figure 2B); GFP-tagged LRRK2 is still competent to bind p62 (Figure 1D). The amino-terminal armadillo and ankyrin repeat domains of LRRK2 are required for interaction with p62. Fragments of p62 (Figure 2C) were expressed in cells expressing GFP-LRRK2 or GFP, and we asked which FLAG-tagged p62 fragments co-precipitated with LRRK2 in GFP immunoprecipitates. We found that the PB1 domain fragment amino acids 1–125 did not interact with LRRK2 when expressed alone; however, fragments of p62 that included amino acids 125–225 did interact with LRRK2. We also found that p62118–440, but not p62167–440, precipitated with LRRK2, indicating the ZZ domain (amino acids 118–167) is necessary for interaction with LRRK2 (Figure 2D). We next sought to determine if this is a direct interaction using recombinant p62 and LRRK2. FLAG-tagged p62118–225 was unstable when expressed in cells, precluding conclusions about interaction using this approach. We did, however, ask if full-length recombinant FLAG-tagged LRRK2 could be isolated with bacterially expressed full-length p62 or the minimal p62-binding domain identified in Figure 2D [p62118–225]. GST or GST-p62 or GST-p62118–225 was incubated with full-length LRRK2, and protein complexes were retrieved with glutathione sepharose and analyzed by immunoblot. We found that LRRK2 was co-precipitated with GST-p62 and GST-p62118-225, but not GST alone (Figure 2E).
Interaction domains of p62/SQSTM1 and LRRK2.
LRRK2 phosphorylates p62 on Thr138
p62 is regulated by phosphorylation by CK2, TBK1/IKKε, p38δ, and ULK1 to regulate its cellular functions. We therefore investigated if LRRK2 not only bound p62, but could also phosphorylate it. Using recombinant p62 and recombinant LRRK2, our initial analysis indicated that LRRK2 could indeed phosphorylate p62. To analyze p62 phosphorylation by LRRK2 and determine the phosphorylation site(s), recombinant p62 was expressed as a GST fusion protein and following purification on glutathione sepharose, the GST tag was removed by cleavage with precision protease. Full-length recombinant LRRK2, purified from mammalian cells as an amino-terminal FLAG-tagged protein, was used to phosphorylate p62 (Figure 3A). Detailed OrbiTrap mass spectrometry analysis of in vitro kinase reaction products revealed several potential sites of phosphorylation. The table in Figure 3 shows the sites of phosphorylation cumulatively identified over several MS determinations (shown in Supplementary Figure S3) with A score and localization probabilities indicated and spectral counts for the number of times the peptide has been observed (Spc). Because LRRK2 binds p62 through its ZZ domain, we further investigated the sites identified within this domain and generated p62 phosphosite mutant T138A, T164A, and S176A recombinant proteins. p62 T138A, T164A, T138A/T164A, S176A, and T138A/T164A/S176A were subjected to in vitro kinase reactions with LRRK2970–2527 (G2019S) recombinant protein. We found decreased LRRK2-mediated phosphorylation when Thr138 was substituted with Ala, while T164A and S176A had no impact (Supplementary Figure S4), indicating that the Thr residue at position 138 is the likely site of LRRK2 modification.
Identification of residues on p62 phosphorylated by LRRK2.
Validation of LRRK2 phosphorylation of p62 Thr138
We generated a phospho-specific antibody against pThr138 to validate the mass spectrometry phosphosite analyses and track p62 modification. This antibody is specific for the phospho-Thr138 peptide and does not detect dephosphopeptide as shown in Supplementary Figure S4B. To test this antibody on p62, we compared p62 with p62 T138A phosphorylation by LRRK2970–2527 (G2019S) recombinant protein in the presence or absence of LRRK2 inhibitor GNE1023 (Figure 3B, left panel) and full-length LRRK21–2527 and kinase-inactive LRRK21–2527 (D1994A) (Figure 3B, right panel). Similar to Supplementary Figure S4, we found that LRRK2 preferentially phosphorylates Thr138. Our antibody specifically recognizes p62 phosphorylated at Thr138 by LRRK2 on WT p62 but neither on p62 T138A nor in reactions where inactive LRRK2 or LRRK2 inhibitor was included. We tracked LRRK2 kinase activity with a pThr1491 antibody, which reveals that LRRK2 autophosphorylation kinase activity was absent in the presence of the kinase-inactive mutant D1994A.
LRRK2 phosphorylates p62 in cells
Thus far, we have found that LRRK2 forms an endogenous complex with p62 in cells and that LRRK2 can phosphorylate p62 in vitro. We next sought to determine if LRRK2 could phosphorylate p62 in cells and if inhibition of LRRK2 would suppress p62 phosphorylation at Thr138. We examined the dose effects of MLi-2 inhibition of LRRK2 in cells expressing GFP-LRRK2 (WT, D2017A and G2019S) and FLAG-p62 (WT and T138A). We found that, similar to the Rab GTPase substrates, concentrations of 3–30 nM MLi-2 decreased p62 phosphorylation at T138A (Figure 4A). We tracked the effects of the potent LRRK2 inhibitor with pSer935 and pSer1292 antibodies; dephosphorylation of the autophosphorylation site (Ser1292) and that of the upstream kinase site (Ser935) were concomitant with p62 dephosphorylation. We observed a similar reduction in p62 phosphorylation using two other structurally diverse inhibitors, PF475 and GNE1023, with IC50 values below 100 nM, compared with an IC50 of below 10 nM for MLi-2 (Supplementary Figure S5A,B). We used MLi-2 on control EBV-transformed lymphoblasts to ask if inhibition of endogenous LRRK2 activity blocks endogenous p62 phosphorylation at pThr138. We found that 10 nM MLi-2 reduces the phospho-Thr138 signal on p62 immunoprecipitates (Figure 4B), demonstrating that p62 is an endogenous substrate of LRRK2. We next compared the rates of dephosphorylation of LRRK2 and p62 pThr138 after inhibition with MLi-2. We found that LRRK2 autophosphorylation is reduced within 10 min, while pSer935 and pThr138 are ablated by 40 min (Figure 4C). To verify whether LRRK2 inhibitor treatment was specific to blocking LRRK2 activity, we employed the A2016T mutation of LRRK2, which shows reduced sensitivity to LRRK2 inhibitors. The treatment of cells expressing p62 and WT LRRK2 with MLi-2 showed reduced p62 phosphorylation but not treatment of cells expressing A2016T with p62 (Figure 4D).
LRRK2 phosphorylates p62 Thr138 in cells.
The dephosphorylation of p62 Thr138 is rapid, revealing a phosphatase activity against this site. To determine if PP1- or PP2-type phosphatases are responsible for p62 dephosphorylation, we employed a pharmacological approach using okadaic acid to inhibit PP2 and Calyculin A to inhibit PP1. Previously, we showed that Calyculin A prevents LRRK2 inhibitor-induced dephosphorylation . Figure 5 shows that inclusion of 100 nM okadaic acid does not affect p62 dephosphorylation in the presence of 10 nM MLi-2, while 20 nM Calyculin A enhances p62 phosphorylation on its own, and also prevents dephosphorylation when co-treated with 10 nM MLi-2 for 30 min, thus implicating PP1-type and not PP2-type phosphatases in the dephosphorylation of LRRK2-phosphorylated p62.
p62 pThr138 is regulated by a Calyculin A-sensitive phosphatase.
A subset of pathogenic PD-associated LRRK2 mutations enhance p62 phosphorylation
Pathogenic PD mutations found in the Roc/COR and kinase domains exhibit increased kinase activity in cells, as measured by autophosphorylation of pSer1292 and phosphorylation of downstream Rab GTPase substrates. As we found that LRRK2 phosphorylates p62 in cells and discovered p62 to be a bona fide substrate of LRRK2, we next investigated if pathogenic PD mutations modulate p62 phosphorylation similar to what has been reported for LRRK2 kinase activity in vitro and against Rabs in cells. To do this, we expressed GFP-LRRK2 (WT, N1437H, R1441G, Y1699C, G2385R and G2019S) and FLAG-p62 WT in cells and analyzed p62 phosphorylation (Figure 6A). Immunoblotting with anti-pThr138 antibody, we found that Thr138 phosphorylation was only observed with active LRRK2 mutants and dramatically reduced in GFP vector control or kinase-inactive LRRK2 (D2017A). LRRK2 N1437H, R1441G, Y1699C and G2019S display increased p62 Thr138 phosphorylation, compared with WT kinase (Figure 6A and quantitated in Figure 6B). We observed increased kinase activity in N1437H, R1441G, Y1699C and G2019S-LRRK2 in cells with the pSer1292 antibody, similar to the increase in p62 phosphorylation. We observed decreased kinase activity of I2020T and G2385R LRRK2 by pSer1292 and p62 pThr138 immunoblotting, which reflects previous analyses of these LRRK2 mutants [90,91]. In line with similar reports, LRRK2 pSer935 is decreased in N1437H, R1441G, Y1699C and G2385R.
Effects of pathogenic PD-associated LRRK2 mutations on p62 phosphorylation in cells and in vitro.
There are discrepancies between the observed increased kinase activity of Roc/COR mutants (e.g. R1441C/G, Y1699C) in vivo and lack thereof in vitro. Increased pSer1292 and Rab phosphorylation is observed in cells, where only slight increases in activity on peptide substrates (Nictide and LRRKtide) or native protein substrates (Rabs) are found in vitro [58,90,91]. We therefore asked if the phosphorylation pattern of p62 was similar in vitro with the same LRRK2 mutants. We used full-length recombinant LRRK2 proteins harboring the R1441G, Y1699C, G2019S and G2385R mutations, and compared phosphorylation of recombinant, tag-free p62 protein and recombinant tag-free Rab8 for in vitro kinase assays. As was observed previously for the Rab substrates , we observed increased phosphorylation with LRRK2 G2019S and diminished phosphorylation with the G2385R mutation, while we observed similar phosphorylation of p62 by R1441G and Y1699C compared with WT protein (Figure 6C). Given that these mutations show similar activity as WT protein in vitro, but increased activity in cells, these results indicate that the increased activity of the Roc/COR mutations observed in cells could be mediated by a cellular factor or event as proposed in ref. .
LRRK2 amino-terminus is required for optimal substrate phosphorylation
We mapped the interaction domain of LRRK2 and p62 to the amino-terminus of LRRK2 and the ZZ domain of p62. We therefore examined whether structural deletions of the LRRK2 interaction domain with p62 would affect regulation of p62. We tested LRRK21-2527, LRRK2970-2527 and LRRK21326-2527, all shown to be active in their recombinant form, for their ability to phosphorylate p62 in cells. We found that deletion of the LRRK2 amino-terminus ablates p62 phosphorylation by LRRK2970-2527 and LRRK21326-2527 proteins (Figure 7A). We next investigated if the amino-terminus is similarly needed to regulate Rab substrates Rab7L1, Rab8 and Rab10. LRRK2 phosphorylation of Rab7L1 tracked similarly to that of p62, where both LRRK2970-2527 and LRRK21326-2527 did not phosphorylate Rab7L1 at Thr71 (Figure 7B, left panel). Interestingly, LRRK2970-2527 was able to phosphorylate both Rab8 and Rab10, though LRRK21326-2527 was unable to modify these Rabs (Figure 7B, middle and right panels). These data further implicate the amino-terminus of LRRK2 in regulation of its kinase activity against certain substrates. Ito et al.  showed that blocking LRRK2 phosphorylation at Ser935 significantly reduced Rab10 phosphorylation; we therefore determined whether regulation of the crucial phosphorylation site Ser910/935 is involved in p62 regulation. To investigate this, we expressed p62 in cells expressing WT, D2017A and S910/935A LRRK2, and found that although S910/935A is active (pSer1292), it does not phosphorylate p62 (Figure 7C).
The LRRK2 amino-terminus and Ser910/935 phosphorylation are required for optimal substrate phosphorylation.
The carboxy-terminus of p62 influences its phosphorylation by LRRK2
A previous report indicates that there is interplay in the regulation of pSer407 and pSer403 in p62. We therefore needed to understand if there could be any impact of the carboxy-terminus or its phospho-regulatory sites on the phosphorylation of p62 by LRRK2. This was accomplished by mutating these sites to alanine or phosphomimetic glutamate and observing if this would alter LRRK2 phosphorylation of Thr138. We saw that neither S403A or S407A alanine substitutions nor the phosphomimetic glutamate substitutions, S403E (which also cross-reacts with anti-pSer403 antibodies) or S407E, significantly altered LRRK2 G2019S phosphorylation on Thr138 (Figure 8A). However, removing the ability of p62 to bind ubiquitin blocked Thr138 phosphorylation. Phe406 in the UBA domain of p62 is crucial for ubiquitin binding, and we found that substitution of Phe406 with Valine, a mutation known to reduce ubiquitin binding to p62 [25,93], decreased the LRRK2 phosphorylation of p62 Thr138 (Figure 8A). Similarly, when we deleted the UBA domain, we also reduced p62 phosphorylation at Thr138 (Figure 8B), despite maintaining its interaction with LRRK2 (Figure 2).
The p62 UBA domain is necessary for LRRK2 phosphorylation at Thr138.
p62 contributes to LRRK2-mediated neuronal cell death
Expression of PD-associated mutants of LRRK2 in primary neuronal cultures induces cell toxicity including neurite shortening and apoptosis [60,94–98]. To determine if p62 contributes to mutant LRRK2-induced neuronal cell toxicity, we expressed WT LRRK2 or mutant LRRK2 (G2019S) with p62 or p62 T138A (Supplementary Figure S6 and Figure 9). We observed intense co-localization of LRRK2 with p62 and condensed nuclei (Figure 9A). The classical apoptotic nuclear morphology co-localized with activated caspase-3 immunostaining (Figure 9A, red). When these cells were scored for the percentage of transfected neurons exhibiting condensed nuclear morphology, we found that G2019S increases the percentage of cells with condensed nuclei that stain positive for caspase 3, over WT LRRK2 or p62 expression alone. However, the LRRK2 unphosphorylatable mutant p62 Thr138Ala fails to enhance the toxicity of mutant LRRK2 (Figure 9B). A representative image of neuronal culture purity as indicated by βIII-Tubulin staining is shown in Figure 9C. Additional representative images of neurons expressing p62 (and LRRK2) and labeled for active caspase-3 are shown in Supplementary Figure S7. Here, we show that co-expression of LRRK2 (G2019S) with p62 significantly increases the number of apoptotic neurons.
p62 participates in mutant LRRK2-induced neuronal death signaling.
Almost 15 years after the LRRK2 locus was found to be associated with PD, only a subset of endogenous substrates have been highly validated in the field [51,58,99], namely the Rab GTPases. In the present study, we contribute another protein to this repertoire of LRRK2 substrates, the signaling adapter p62/SQSTM1. We showed that p62 is an endogenous interactor of LRRK2. Using mass spectrometry and mutational analysis, we found that LRRK2 phosphorylates p62 on Thr138 in vitro. Using a specific pThr138 antibody that readily detects p62 phosphorylated on Thr138, we then showed that p62 phosphorylation is diminished in cells treated with specific and selective LRRK2 inhibitors and that PD-associated mutations in LRRK2 (N1437H, R1441G, Y1699C and G2019S) increase p62 phosphorylation. The precise mechanism of how p62 Thr138 phosphorylation alters its downstream function in established pathways is still to be investigated. However, we report here that when p62 and mutant LRRK2 are introduced together in primary neurons, a synergistic lethality is observed with increased apoptosis (Figure 9), which is rescued by mutation of Thr138 to Ala.
We originally observed p62 association with LRRK2 in a targeted screen of multiple putative interacting chaperone and adapter proteins. Hsp27 provides thermotolerance and stabilization of denatured protein for remedy by refolding chaperones [100,101], and interacts with LRRK2. We and others have previously identified Hsp90 and Cdc37 interacting with LRRK2 [86,87,102–105]. Hsp70 and Hsc70 are also known LRRK2-interacting proteins . Expression of Hsp110, a protein disaggregase, did not interact with LRRK2. HDAC6 and p62 facilitate the removal of aggregated and ubiquitinated proteins and, interestingly, these proteins specifically co-immunoprecipitated with LRRK2. LRRK2 was previously reported to interact with acetylated microtubules, which are produced by enhanced HDAC6 activity , supporting LRRK2 as playing a regulatory role for HDAC6. Endogenous LRRK2 interacts with endogenous p62, and the specificity of this interaction is demonstrated by the more selective co-immunoprecipitation with LRRK2 instead of LRRK1 (Figure 1). The domain structure is very similar between LRRK2 and LRRK1, except for divergent amino-termini. This probably explains the identification of the LRRK2 amino-terminus as the interaction domain with p62 (Figure 2). We additionally found that p62 binds to the LRRK2 armadillo domain through its ZZ domain (p62118–225), which is similar to what was seen in Park et al. . There is overlap in the conclusion but a slight discrepancy in the results, which could be due to the use of internal deletion constructs  versus expression-tagged fragments used here. Typically, we observe an enhanced interaction of LRRK2 and p62 in the presence of LRRK2 inhibitors; however, this may be context-dependent or transient.
Some kinases interact with their substrates stably , and this is true for LRRK2 and p62. We mapped Thr138 on p62 as a specific site of LRRK2 phosphorylation, and we developed phospho-specific antibodies against this site. Interestingly, a tryptic peptide from p62 that encompasses the Thr138 site has been identified in multiple studies [108–111] but with di-Gly linkage on the Lys at the +3 position from Thr138; this could have precluded or complicated identification of pThr138 in unbiased phospho-proteomic screens and increased our technical difficulty in detecting endogenous pThr138. We established a dependence of LRRK2 activity for phosphorylation of p62 Thr138 in cells using three structurally distinct inhibitors. Using the specific inhibitor MLi-2, LRRK2 phosphorylation of p62 was blocked within 40 min and was restored with the inhibitor-resistant mutant (A2016T). Importantly, we found that inhibition of endogenous LRRK2 decreased p62 phosphorylation on Thr138 in lymphoblasts (Figure 4B). Having established that blocking LRRK2 kinase activity results in decreased p62 phosphorylation, we next characterized the in vivo phosphatase activity on p62. The downstream target p62 was dephosphorylated at a similar IC50, but at a more rapid rate than pSer1292 (Figure 4). The rapid dephosphorylation of p62 indicates that a phosphatase actively regulates p62. The LRRK2 signaling pathway has been implicated to involve PP1 and possibly PP2A  in the role of Ser935 dephosphorylation. Using a pharmacological approach, we narrowed the class of phosphatases involved in p62 Thr138 dephosphorylation to PP1-type enzymes, and possibly PP2A (Figure 5).
We found that pathogenic PD-associated LRRK2 mutants (N1437H, R1441G, Y1699C and G2019S) enhance the phosphorylation of p62 in cells. However, this enhancement was only observed with G2019S in vitro, compared with the R1441G and Y1699C mutants. This reflects similar modification patterns observed with the recently reported Rab substrates comparing in vitro with in vivo activity of LRRK2 . This could be due to cellular factors or localization differences in cells that are not present in vitro that contribute to the regulation of LRRK2 kinase activity. The risk factor mutation G2385R, which was reported to have decreased activity and reduced Ser935 phosphorylation , exhibited reduced phosphorylation of p62 in cells and in vitro, similar to the Rab substrates (Figure 6). We observed that the I2020T mutation did not increase LRRK2 phosphorylation of p62, raising important distinctions in biochemical changes caused by different disease or risk-enhancing mutations. These data indicate that, similar to pSer935 and phospho-Rab substrates, p62 pThr138 is a viable pharmacodynamic marker and activity marker for LRRK2.
There is yet to be a crystal structure of full-length LRRK2; however, a molecular model, derived from cryo-EM and cross-linking studies, provides useful insights into the general organization of the dimer , which other structural studies in general support [114,115]. The amino-terminus appears to contact the Roc–COR domain interface to regulate activity. In support of this, Ito et al.  determined that phosphorylation of LRRK2 pSer910/935 is necessary for Rab 10 phosphorylation. In the context of the novel substrate p62, we found that pSer910/935 is also necessary for p62 Thr138 phosphorylation. Using LRRK2 amino-terminal deletion mutants, we showed that the amino-terminus influences phosphorylation of p62 and Rab substrates. It is therefore likely that phospho-regulation of the LRRK2 amino-terminus (e.g. Ser910/935/955/973) indirectly regulates LRRK2 substrate phosphorylation, and therefore the kinases and phosphatases that modify LRRK2 are prime targets for elucidation. This is distinct from in vitro conditions (Figure 3), where LRRK2970-2527 is able to phosphorylate p62, but it is, however, further reflective of the differences in LRRK2 kinase activity observed in vitro versus in cells as is seen with Roc/COR mutants. This is also the case for the Rab proteins, where LRRK2970-2527 is able to phosphorylate Rabs in vitro , but pSer910/pSer935 is required in cells .
There are consistent reports of LRRK2 modulation of the autophagic process. However, the mechanistic role of LRRK2 in regulating autophagy has been elusive to date, i.e. no members of the autophagic machinery (i.e. ATG proteins) have been shown to be validated LRRK2 kinase substrates. p62 is phosphorylated by multiple kinases throughout several domains to regulate its function and many of these phosphorylation events have been shown to have interdependent regulation. We show that the carboxy-terminal ubiquitin-binding domain of p62 influences the phosphorylation of the ZZ domain by LRRK2. It is therefore possible that the ubiquitin-binding function of p62, a crucial aspect of its autophagic role, feeds into LRRK2 regulation of p62 and thus autophagy. Therefore, the ubiquitin-binding function of p62 bridges LRRK2 activity to autophagy regulation, where the impact of hyperphosphorylation of p62 will need to be investigated.
Our data linking p62 to LRRK2 PD support previous reports of p62 involvement with aggregated αSyn protein inclusions in PD as well as in other neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS), polyglutamate disorders and multiple system atrophy (MSA) [3–11]. p62 binds to synuclein inclusions and pathology increases in the contralateral hemisphere from fibrillar αSyn injection . This places p62 intracellularly with endogenous αSyn and also inclusions in cells affected by synuclein through uptake from the extracellular milieu. Interestingly, cytoplasmic aggregates of LRRK2 expressed in cell culture and primary neurons co-localize with p62 . Furthermore, p62 knockout in an αSyn transgenic mouse has been found to increase Lewy body-like pathology .
p62 is involved in multiple cellular processes that could indeed intersect with reported functions of LRRK2, from mitochondrial maintenance to vesicular trafficking [12,13,15,18,46,62,63,65,118–125]. For LRRK2, expression of PD-associated mutants in cultured neurons leads to enhanced cell toxicity [61,94,95]. When we investigated if p62 modulated this phenotype, we indeed found a synergistic toxic effect of mutant LRRK2 and p62 expression, dependent on Thr138 (Figure 9). It is possible that, in this context, overexpressed hyperphosphorylated p62, co-expressed with G2019S-LRRK2, fails to activate the Nrf2 cytoprotective pathway similar to that recently reported in cultured neurons in which elevated levels of phosphorylated p62 (at Ser351) are observed following proteasome inhibition . In fact, it was recently shown that up-regulating Nrf activity decreased LRRK2 neuronal toxicity . We have identified a co-operative role of p62 in LRRK2 kinase activity-dependent toxicity, but delineating the effects on other p62 functions is the topic of future studies. The interplay of p62 phosphorylation on disease pathogenesis is only now being uncovered and, indeed, there are other LRRK2-influenced phosphorylation sites , further linking the phospho-regulation of p62 to PD. In conclusion, this work identifies p62 as a novel substrate of LRRK2. With this novel substrate, we validate the necessity of the amino-terminus for LRRK2 substrate phosphorylation and uncover p62 as a potential mediator of PD-mutant LRRK2 toxicity. [127,128]
casein kinase 1
false discovery rates
full length at half maximum
Kelch-like ECH associated protein
leucine-rich repeat kinase 2
mammalian target of rapamycin complex 1
proximity ligation assay
TGF-beta activated kinase
Unc-51 like kinase 1
R.J.N. conceived and designed the study. J.Z. originally observed the LRRK2–p62 interaction. A.F.K., J.Z., M.F.B., A.M., S.N., T.P.M., W.H.W., H.J.R. and R.J.N. conducted and analyzed the experiments. A.F.K., H.J.R. and R.J.N. wrote the manuscript.
The authors thank the Brin/Wojcicki Foundation and the Michael J. Fox Foundation for funding this work.
We thank Troomonos Mou for technological assistance. We are grateful to Dario Alessi for providing MLi-2, Rab antibodies and plasmid reagents. We are also grateful to the many plasmid depositors from Addgene.
The Authors declare that there are no competing interests associated with the manuscript.
Present address: MRC Protein Phosphorylation and Ubiquitylation Unit, University of Dundee, Dundee, U.K.