Mutations in the LRRK2 (leucine-rich repeat kinase 2) gene have been identified in PARK8, a major form of autosomal-dominantly inherited familial Parkinson's disease, although the biochemical properties of LRRK2 are not fully understood. It has been proposed that LRRK2 predominantly exists as a homodimer on the basis of the observation that LRRK2, with a theoretical molecular mass of 280 kDa, migrates at 600 kDa (p600 LRRK2) on native polyacrylamide gels. In the present study, we biochemically re-examined the nature of p600 LRRK2 and found that p600 LRRK2 was fractionated with a single peak at ~272 kDa by ultracentrifugation on a glycerol gradient. In addition, p600 LRRK2 behaved similarly to monomeric proteins upon two-dimensional electrophoretic separation. These results suggested a monomeric composition of p600 LRRK2 within cells. The p600 LRRK2 exhibited kinase activity as well as GTP-binding activity, and forced dimerization of LRRK2 neither upregulated its kinase activity nor altered its subcellular localization. Collectively, we conclude that the monomer form of LRRK2 is predominant within cells, and that dimerization is dispensable for its enzymatic activity.
PD (Parkinson's disease) is one of the most common neurodegenerative diseases of adult-onset affecting the extrapyramidal motor system . The characteristic neuropathological changes in the brains of patients with PD include neuronal loss of the brain stem monoaminergic nuclei, e.g. dopaminergic neurons in substantia nigra or noradrenergic neurons in locus caeruleus, accompanied by formation of Lewy bodies in the remaining neurons [2–4]. Most PD patients develop the disease in a sporadic manner, whereas a subset of patients inherit PD as autosomal dominant or recessive traits [FPD (familial PD)]. At least six genes have so far been identified as the responsible genes for FPD; among these, LRRK2 (leucine-rich repeat kinase 2) has been identified as the causative gene for PARK8, an adult-onset autosomal-dominant form of FPD [5–7]. Six mutations (R1441C/G/H, Y1699C, G2019S and I2020T) in the LRRK2 gene have been shown to segregate with the clinical manifestation in PARK8 . LRRK2 mutations have also been found in sporadic PD cases from some ethnic populations , and genetic variation around the LRRK2 locus has been identified to modulate the risk for sporadic PD by two independent genome-wide association studies [10,11], leading to the notion that LRRK2 is a major genetic factor involved in PD. It has been reported that clinical manifestations of PARK8 are similar to those of typical sporadic PD patients, and that a somewhat pleiomorphic neuropathological feature frequently associated with α-synuclein deposition, including cases with pure nigral degeneration without fibrous protein deposits or with tau-positive inclusions, characterizes the disease . Therefore analysis of the pathomechanism whereby mutations in the LRRK2 gene cause neuronal degeneration would provide important clues to the pathogenesis of familial as well as sporadic forms of PD.
LRRK2 is a large cytoplasmic protein composed of 2527 amino acids that harbours several functional domains: leucine-rich repeat and WD40 repeat motifs, which are presumed to be involved in protein–protein interactions; a ROC (Ras of complex proteins) domain harbouring the GTP-binding activity; and a protein kinase domain. The kinase activity of LRRK2 has been shown to be up-regulated by FPD mutations (especially by G2019S mutation) [13,14]. Also, it has been reported that overexpression of the FPD mutant LRRK2 induces cell death in a manner dependent on its kinase activity [15,16]. Accordingly, an abnormal activation of LRRK2 by genetic mutations or other factors is implicated in the mechanism of neurodegeneration in PD.
Dimerization and complex formation are often involved in the regulation of the activity of protein kinases. For example, ATM (ataxia telangiectasia mutated), a large cytoplasmic kinase, is known to form a homodimer under normal conditions, whereas dissociation of the homodimer occurs upon DNA damage, resulting in the activation of the kinase . LRRK2 has also been reported to form a homodimer on the basis of the finding that differentially tagged LRRK2 could be co-immunoprecipitated upon overexpression in cultured cells . Although the direct elucidation of dimerization by analytical ultracentrifugation has not been accomplished because of the difficulties in purifying a sufficient amount of full-length LRRK2, the observation that the apparent molecular mass of native LRRK2 was estimated to be ~600 kDa by gel filtration and native-PAGE, which is almost double the theoretical mass of LRRK2 (~280 kDa), provided a strong support for the homodimer formation of LRRK2 [18,19]. Sen et al.  reported that existence of the apparently dimeric substance is correlated with the active state of LRRK2. However, the amount of LRRK2 homodimer was quite small in the co-immunoprecipitation experiment, consistent with a recent report by Berger et al. , suggesting that the dimer form of LRRK2 is a minor subspecies in cells. These results prompted us to carefully examine whether a dimer or monomer is the predominant and kinase-active form of LRRK2. In the present study we show that LRRK2 migrating at the ~600 kDa position on native-PAGE represents a monomer of LRRK2 by a series of biochemical experiments. We also revealed that the monomer form of LRRK2 possesses kinase activity as well as GTP-binding activity. Furthermore, chemical-induced dimerization neither activated the kinase activity nor altered the subcellular localization of LRRK2. These results suggested that LRRK2 predominantly exists as a monomer within cells, and that dimerization is dispensable for its basal kinase activity.
MATERIALS AND METHODS
Construction of expression plasmids
The expression plasmids encoding full-length human LRRK2 (wild-type, K1347A, T1348N and K1906M) cloned into the p3×FLAG-CMV-10 vector (Sigma) were constructed as described previously . The expression plasmid of ΔN-LRRK2 (residues 1326–2527) cloned into the pDEST27 vector (Invitrogen) was constructed as described previously . The expression plasmid of the Fv [FKBP (FK506-binding protein) variant] (pC4-Fv1E) was obtained from Ariad Pharmaceuticals. DNA fragments encoding full-length LRRK2 or ΔN-LRRK2 (1326–2527) with a C-terminal 2×Myc tag were cut and ligated into pC4-Fv1E. All constructs generated from PCR products were verified by DNA sequencing.
Cell culture and transfection
HEK (human embryonic kidney)-293 cells and 3T3-Swiss albino (Swiss 3T3) cells were maintained in DMEM (Dulbecco's modified Eagle's medium; Sigma) supplemented with 10% fetal bovine serum and 100 units/ml penicillin/100 μg/ml streptomycin (Invitrogen) at 37°C in 5% CO2 atmosphere. Transient expression in HEK-293 cells was performed by transfecting the plasmids using FuGENE6™ (Roche) according to the manufacturer's instructions. ON-TARGET plus siRNAs (small interference RNAs) targeting murine LRRK2 or GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and non-targeting siRNAs were purchased from Dharmacon. RNA interference in Swiss 3T3 cells were performed by transfecting the siRNAs using Lipofectamine™ RNAiMAX (Invitrogen) according to the manufacturer's instructions.
Antibodies and immunochemical analysis
A rabbit polyclonal antibody raised against the C-terminus of human LRRK2 (residues 2500–2527) was purchased from Novus Biologicals (NB300-268). Anti-FLAG M2 antibody was purchased from Sigma. Anti-Myc (9B11), anti-GAPDH, anti-mTOR (mammalian target of rapamycin), anti-raptor (regulatory associated protein of mTOR) and anti-ATM antibodies were purchased from Cell Signaling Technology. For immunoprecipitation of the 3×FLAG-tagged protein, 48 h after transfection, transfected cells were lysed in lysis buffer [50 mM Tris/HCl (pH 7.4), 150 mM NaCl, 0.5% Nonidet P40, 20 mM MgCl2, Complete™ protease inhibitor cocktail EDTA-free (Roche) and PhosSTOP phosphatase inhibitor cocktail (Roche)] for 30 min at 4°C. Anti-FLAG M2 antibody and Protein G–Sepharose 4 Fast Flow (GE Healthcare) were added to the lysate after preclearing with CL4B Sepharose (GE Healthcare) in the absence of antibodies, and the mixture was incubated for 2 h at 4°C. Precipitated immunocomplexes were then washed in the lysis buffer five times and subjected to immunoblot analysis or to an in vitro analysis. For immunoblot analysis, precipitated proteins were solubilized by boiling in SDS/PAGE sample buffer [2% SDS, 80 mM Tris/HCl (pH 8.0), 15% glycerol and 1% 2-mercapthoethanol] and analysed by immunoblotting as described previously . Ponceau S staining was carried out by soaking membranes in the staining solution [0.1% Ponceau S (Sigma) and 5% acetic acid] for 5 min and washing the membranes in distilled water. In some samples, gels were soaked into Colloidal Blue staining solution (Invitrogen) and stained according to the manufacturer's instructions. For pull-down of GST (glutathione transferase)-fusion proteins, glutathione–Sepharose 4B (GE Healthcare) was used instead of the anti-FLAG antibody and Protein G–Sepharose beads. Detection of in vivo phosphorylation of LRRK2 and the in vitro kinase assay were performed as described previously .
BN-PAGE (blue native PAGE) and 2D (two-dimensional) BN-PAGE/SDS/PAGE analysis
BN-PAGE was performed according to the manufacturer's instructions (Invitrogen). Briefly, cells were harvested and resuspended in NativePAGE™ 4× sample buffer (Invitrogen) and lysed using four cycles of freezing and thawing in the buffer. Protein concentration was determined by BCA (bicinchoninic acid) protein assay kit (Pierce), and 10 μg of protein was applied to NativePAGE™ 4–16% gels along with NativeMark unstained protein standard (Invitrogen). After electrophoresis, gels were soaked into the blotting buffer [25 mM Tris, 195 mM glycine and 10% methanol] containing 0.1% SDS for 15 min. Subsequent immunoblotting was performed as described above. For normal denaturation of samples, cell lysates were added to the indicated concentration of SDS and 2-mercapthoethanol and boiled for 5 min. To induce acid hydrolysis of proteins, cells were resuspended in a harsh denaturing buffer [50 mM Tris/HCl (pH 6.8), 50 mM KCl, Complete™ protease inhibitor cocktail and PhosSTOP phosphatase inhibitor cocktail] and lysed using four cycles of freezing and thawing . Cleared lysates were added to 2% SDS and heated to 110°C for the indicated time. For silver staining of BN-PAGE gels, gels were destained after BN-PAGE using destaining solution [50% methanol and 10% acetic acid] for 12 h, and stained with 2D-Silver Stain II (Daiichi) according to the manufacturer's protocol. 2D BN-PAGE/SDS/PAGE analysis was conducted as follows: lanes of BN-PAGE gels were cut and soaked in SDS/PAGE sample buffer for 30 min on a rotary shaker, and put on the NuPAGE Novex 4–12% Bis-Tris gel [IPG (immobilized pH gradient) well]. After electrophoresis, gels were subjected to immunoblotting or Colloidal Blue staining as described above.
Separation of LRRK2 on a glycerol gradient
Cells were harvested and lysed in lysis buffer [50 mM Hepes/NaOH (pH 7.5), 150 mM NaCl, 2 mM DTT (dithiothreitol), PhosSTOP phosphatase inhibitor cocktail and Complete™ protease inhibitor cocktail] using four cycles of freezing and thawing. Linear glycerol gradients (9–35%) were made using gradient buffer [9% or 35% glycerol, 25 mM Hepes/NaOH (pH 7.5), 1 mM DTT and 0.1× Complete™ protease inhibitor cocktail]. Cleared lysate was applied on top of the gradient and separated as described previously .
Pull-down of LRRK2 by GTP–agarose
GTP–agarose was purchased from Innova Biosciences. The beads were treated with 0.01% BSA in Tris-buffered saline (50 mM Tris/HCl, pH 7.6, and 150 mM NaCl) for 1 h at 4°C before use. After blocking, the beads were equilibrated with lysis buffer [50 mM Tris/HCl (pH 7.5), 50 mM KCl, 5 mM MgCl2, 1 mM EDTA, 0.1 mM DTT and 0.5% digitonin] for 30 min at 4°C. Cells were harvested and lysed in lysis buffer containing Complete™ protease inhibitor cocktail, and the protein concentration was determined using the BCA protein assay kit. Then, 70 μl of 50% GTP–agarose was added to 300 μg of protein and incubated for 2 h at 4°C. The beads were washed with lysis buffer three times and bound proteins were eluted by incubating the beads in 100 μl of elution buffer [10 mM GTP and 1x NativePAGE sample buffer] for 1 h at 4°C. Eluted samples were analysed by BN-PAGE and SDS/PAGE followed by immunoblotting.
Chemical-induced dimerization of LRRK2
The bifunctional compound AP20187, for dimer induction, was obtained from Ariad Pharmaceuticals . HEK-293 cells were transfected with plasmids encoding LRRK2 tagged with Fv at the N-terminus. Transfected cells were treated with 50 nM AP20187 for 4 h, harvested and subjected to BN-PAGE analysis as described above. For analysis of in vivo phosphorylation, transfected cells were treated with 50 nM AP20187 and labelled with [32P]Pi (PerkinElmer) for 4 h and analysed as described previously . AP20187 was included in buffers throughout the analysis to maintain the induced dimers.
BN-PAGE analysis of LRRK2
To estimate the molecular mass of LRRK2 in its native form, we employed BN-PAGE analysis (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/441/bj4410987add.htm). HEK-293 cells transfected with full-length human LRRK2 harbouring a 3×FLAG tag at the N-terminus were lysed as described in the Materials and methods section. Immunoblot analysis of cell lysates following BN-PAGE revealed that overexpressed 3×FLAG–LRRK2 migrated at the ~600 kDa position (p600) associated with smearing substances in HMM (higher molecular mass) ranges, whereas it migrated at the ~280 kDa position on SDS/PAGE gels (Figure 1A). We confirmed that endogenous LRRK2 from mouse Swiss 3T3 cells migrated in a similar manner to overexpressed human LRRK2 on BN-PAGE and SDS/PAGE gels (Figure 1B and Supplementary Figure S2 at http://www.BiochemJ.org/bj/441/bj4410987add.htm). These results are in agreement with previous reports [18–21] as well as with the notion that full-length LRRK2, with a calculated mass of ~280 kDa, forms a dimer in its native form. Assuming that LRRK2 forms a dimer under native conditions, it could be dissociated into monomers under denaturing conditions. Unexpectedly, however, 3×FLAG–LRRK2 remained at an ~600 kDa position on BN-PAGE after treatment of lysates with 1% SDS and boiling, whereas the smearing substance in the HMM ranges disappeared upon denaturation (Figure 1A). A small portion of the HMM substances remained after denaturation (Figure 1A; **), which disappeared by reduction with 2-mercaptoethanol, suggesting that the HMM substances contain the oxidized (e.g. disulfide cross-linked) form of LRRK2. Taken together, these results raised the possibility that the p600 LRRK2 on BN-PAGE gels does not represent a dimer, but a monomer form of LRRK2.
BN-PAGE analysis of full-length LRRK2
2D BN-PAGE/SDS/PAGE analysis of LRRK2
Next, we employed 2D-PAGE to investigate whether the p600 LRRK2 consists of full-length LRRK2. The lysates of HEK-293 cells overexpressing 3×FLAG–LRRK2 were first separated by BN-PAGE and then by SDS/PAGE. Immunoblot analysis revealed that the p600 LRRK2 as well as the HMM smearing substances migrated at an ~280 kDa position on the second-dimension SDS/PAGE gel (Figure 2A), indicating that both p600 and HMM smears consisted of full-length LRRK2. When we denatured the lysate prior to 2D-PAGE, only p600, which consisted of full-length LRRK2, was observed (Figure 2B). To compare the behaviour of LRRK2 on 2D-PAGE gels with those of other monomeric proteins, we stained the 2D-PAGE gel replica after separation of the denatured lysate with Coomassie Brilliant Blue. A large number of spots shaping a zone surrounded by two diagonally hyperbolic curves were observed as reported previously  (Figure 2C); this zone, which we designate the ‘monomer zone’, was considered to consist of the monomer form of proteins, since protein complexes in lysates were disrupted by SDS treatment and boiling (see Supplementary Figures S3A and S3B at http://www.BiochemJ.org/bj/441/bj4410987add.htm). Upon overlaying the immunoblot of LRRK2 to the Coomassie Brilliant Blue-stained gel replica, we confirmed that the p600 LRRK2 was located within the ‘monomer zone’ (Figure 2D). These results further supported the view that the p600 LRRK2 corresponds to a monomer form of LRRK2.
2D BN-PAGE/SDS/PAGE analysis of LRRK2
Prolonged SDS treatment caused degradation of LRRK2
Given that the p600 LRRK2 may correspond to a monomer form of LRRK2, we hypothesized that the ~300 kDa LRRK2-positive band separated on BN-PAGE, frequently described as a LRRK2 ‘monomer’ in the literature [18–21], may represent degradation products of LRRK2. To test this hypothesis, our aim was to cause artificial degradation of LRRK2 in vitro and to examine the migration pattern of the degraded proteins by BN-PAGE. Artificial degradation of LRRK2 was elicited by treating cell lysates with 2% SDS and heating to 110°C for up to 30 min, a condition which has been shown to cause protein degradation during sample preparation for SDS/PAGE . After treating cell lysates in the harsh conditions described above, we observed the appearance of a faint band migrating at ~300 kDa separated on BN-PAGE gel (p300 LRRK2; arrow in Figure 3A). Then we subjected the BN-PAGE gel to 2D-PAGE analysis and found that the p300 LRRK2 migrated at ~200 kDa by the second-dimension SDS/PAGE (Figure 3B; white arrowhead). These results indicated that the p300 LRRK2 corresponds to a degraded form of LRRK2, not a LRRK2 holoprotein. In addition, the p300 LRRK2 lacked the epitope of an anti-LRRK2 antibody that recognizes the C-terminus of LRRK2 (see Supplementary Figure S4 at http://www.BiochemJ.org/bj/441/bj4410987add.htm), supporting the notion that the p300 LRRK2 corresponds to a degraded form of LRRK2.
Prolonged SDS/heat treatment of LRRK2
BN-PAGE analysis of HMM proteins
Since LRRK2 is an HMM protein, we hypothesized that soluble HMM proteins in general exhibit slower migration on BN-PAGE gels compared with their theoretical mass. To examine this assumption, we investigated the migration patterns of several HMM proteins in a monomeric form by BN-PAGE. We first analysed ATM, a well-known HMM kinase with a calculated mass of ~330 kDa (3056 amino acids) . ATM forms a homodimer under normal conditions, whereas it dissociates into monomers under stress conditions (e.g. irradiation) . When endogenous ATM of HEK-293 cells cultured under normal conditions was analysed by BN-PAGE, ATM migrated at ~800 kDa with a faint band at ~600 kDa (see Supplementary Figure S5A at http://www.BiochemJ.org/bj/441/bj4410987add.htm), whereas ATM migrated solely at ~600 kDa upon denaturation and reduction (Supplementary Figure S5A). These results suggested that the bands migrating at ~800 kDa and ~600 kDa represent the dimeric and monomeric forms of ATM respectively. The spot corresponding to the putative monomeric and dimeric ATM was located within and below the ‘monomer zone’ on 2D-PAGE gel respectively (Supplementary Figures S3 and S5B), indicating that the monomer form of the authentic dimeric kinase behaves similarly to the p600 LRRK2. Next, we analysed mTOR, another well-known protein kinase physiologically incorporated into protein complexes, i.e. mTORC1 (mTOR complex 1) and mTORC2 . Without denaturation, endogenous mTOR of HEK-293 cells, with a calculated mass of ~280 kDa (2549 amino acids), migrated at ~900 kDa with a faint band at ~600 kDa after separation by BN-PAGE (Supplementary Figure S5C). In addition, an essential component of mTORC1, raptor, with a calculated mass of ~150 kDa (1335 amino acids), also migrated at ~900 kDa with a faint band at ~300 kDa (Supplementary Figure S5C). These results suggested that the band migrating at ~900 kDa (arrowhead in Supplementary Figure S5C) included mTORC1 consisting of mTOR, raptor and other components. Upon denaturation and reduction, the ~900 kDa band disappeared, and monomeric mTOR and raptor were observed at ~600 kDa and ~300 kDa respectively (Supplementary Figure S5C). Subsequent 2D-PAGE analysis revealed that the spot corresponding to monomeric mTOR was located within the ‘monomer zone’ on the 2D-PAGE gel (Supplementary Figures S5D and S5E). Taken together, these results suggested that the monomer form of HMM proteins, including ATM, a physiologically dimeric kinase, as well as mTOR, a kinase physiologically forming protein complexes, have a tendency to exhibit an apparently slower migration that is deviated from the theoretical masses on BN-PAGE gels, as observed in LRRK2.
Separation of LRRK2 by glycerol gradients
To investigate further whether p600 represents a monomeric or dimeric form of LRRK2, we utilized glycerol velocity gradient centrifugation, an alternative method to separate native proteins depending on their molecular mass. The dependency of separation on the molecular mass of proteins was verified using protein standards and endogenous mTORC1 (see Supplementary Figures S6A and S6B at http://www.BiochemJ.org/bj/441/bj4410987add.htm). Analysis of overexpressed LRRK2 in HEK-293 cells as well as of endogenous LRRK2 from Swiss 3T3 cells by separation on glycerol gradients followed by BN-PAGE revealed that the p600 LRRK2 was fractionated with a single peak around the ~272 kDa fraction (Figures 4A and 4B). Since the apparent molecular mass of ~272 kDa is consistent with the theoretical mass of the monomeric form of LRRK2 (~280 kDa), this result again supported our hypothesis that the p600 LRRK2 corresponds to monomeric LRRK2.
Separation on glycerol gradients and BN-PAGE analysis of LRRK2
Monomeric LRRK2 possesses kinase activity as well as GTP-binding activity
Since LRRK2 harbours both kinase and GTP-binding activities, we then investigated whether monomeric LRRK2 possesses these activities. Overexpressed 3×FLAG–LRRK2 in HEK-293 cells was first fractionated on a glycerol gradient and then subjected to an in vitro kinase assay following immunoprecipitation. Wild-type LRRK2, but not a K1906M mutant which lacks the kinase activity, exhibited autophosphorylation activity with a single peak around the ~272 kDa fraction (Figures 5A and 5B). This result suggested that monomeric LRRK2 harbours the kinase activity. We then tested the GTP-binding activity of LRRK2 by pull-down of the cell lysates with GTP–agarose beads. Lysates of HEK-293 cells overexpressing 3×FLAG–LRRK2 were incubated with GTP–agarose beads and bound LRRK2 was eluted with 10 mM GTP and analysed using BN-PAGE and SDS/PAGE. The T1348N mutant LRRK2 lacking GTP-binding activity was not pulled down by GTP–agarose beads as expected. Analysis of the eluted samples by BN-PAGE revealed that the p600 LRRK2 was bound to the GTP-agarose beads (Figure 5C), suggesting that the monomeric LRRK2 harbours GTP–binding activity.
The autophosphorylation and GTP-binding activity of LRRK2 separated by glycerol gradients
Chemical-induced dimerization of LRRK2 failed to alter LRRK2 activity
Although LRRK2 was shown to exist predominantly as a monomer in the resting state, a small portion of LRRK2 was suggested to exist as a dimer on the basis of the co-immunoprecipitation experiment (see Supplementary Figure S7 at http://www.BiochemJ.org/bj/441/bj4410987add.htm). Therefore we thought that the functional consequences of dimerization would be worth investigating, since several kinases are known to be activated by changing their oligomerization states in response to upstream events [17,29,30]. We thus analysed the effect of forced dimerization on the kinase activity of LRRK2 using a chemical-induced dimerization system [26,31]. In this experimental paradigm, full-length LRRK2 was fused with modified Fv at the N-terminus, which forms a dimer upon treatment with a bifunctional small compound, AP20187, together with a C-terminal 2×Myc tag (Figure 6A). When cells overexpressing Fv–LRRK2–2×Myc were treated with AP20187, the p600 LRRK2 on BN-PAGE gels was shifted upward, indicating a dimer formation (Figure 6A). We verified further the induction of LRRK2 dimerization by showing that co-immunoprecipitation of differentially tagged LRRK2 (Fv–LRRK2–2×Myc and Fv–LRRK2–3×FLAG) was markedly increased upon treatment with AP20187 (see Supplementary Figure S8A at http://www.BiochemJ.org/bj/441/bj4410987add.htm). When we examined the kinase activity of Fv–LRRK2 immunoprecipitated from cells treated with AP20187 using an in vitro kinase assay, the kinase activity on GST–LRRKtide , as well as the autophosphorylation activity, was not altered upon treatment with AP20187 (Figure 6B), suggesting that dimerization does not cause up-regulation of the intrinsic kinase activity of LRRK2. As previously described by Berger et al. , dimerization of LRRK2 has been implicated in its membrane localization. Taking advantage of our forced dimerization system, we investigated whether dimerization is sufficient for the membrane localization of LRRK2. Biochemical fractionation revealed that ~20% of overexpressed Fv–LRRK2 was fractionated into the membrane fraction (Supplementary Figure S8B and Supplementary Materials and methods at http://www.BiochemJ.org/bj/441/bj4410987add.htm), suggesting membrane association of LRRK2 as reported previously [13,21]. However, neither biochemical fractionation nor immunocytochemical analysis revealed the alteration of the subcellular localization of Fv–LRRK2 upon treatment with AP20187 (Supplementary Figures S8B and S8C). These results suggested that dimerization is dispensable both for the intrinsic activity and membrane localization of LRRK2.
Chemical-induced dimerization of LRRK2
Dimerization is often involved in the regulation of enzymatic activities as well as the biochemical properties of various types of proteins, among which kinases are the most classical examples. LRRK2 has been shown to self-interact in yeast two-hybrid screening , as well as in co-immunoprecipitation experiments , suggesting that LRRK2 also forms a dimer in cells. Moreover, it has also been shown that full-length LRRK2 predominantly migrates at ~600 kDa on BN-PAGE gels [18–20], implying that LRRK2 predominantly exists as a dimer in cells. However, we and others have shown that the dimeric form of LRRK2 does not exist abundantly in cells by co-immunoprecipitation experiments (Supplementary Figure S7) . To clarify this discrepancy, we carefully examined the ~600 kDa substance of LRRK2 by multiple lines of analytical techniques and found that: (i) LRRK2 migrating at ~600 kDa on BN-PAGE gels represents the monomer form of LRRK2; (ii) the monomeric form of LRRK2 possesses both kinase and GTP-binding activities; and (iii) induced dimer formation of LRRK2 in cultured cells does not affect its intrinsic kinase activity or its subcellular localization.
We first confirmed that LRRK2 predominantly migrated at ~600 kDa (p600) on BN-PAGE gels as reported previously (Figure 1) [18–21]. Although the apparent molecular mass of the p600 LRRK2 on BN-PAGE gels is consistent with the theoretical mass of a dimer, we concluded that the p600 LRRK2 represents the monomer form for the following reasons. First, the p600 LRRK2 was not diminished, but rather intensified, under normal denaturing conditions (0.5–2% SDS, reducing agent, boiling for 5 min) (Figure 1A), in which usual protein complexes are expected to dissociate. In fact, protein homodimers (e.g. ATM) as well as protein complexes (e.g. mTORC1) were dissociated into lower molecular mass substances upon denaturation (Supplementary Figures S3 and S5). Secondly, the spot of p600 LRRK2 was localized within the ‘monomer zone’ on 2D BN-PAGE/SDS/PAGE gels (Figure 2). Since Camacho-Carvajal et al.  reported that proteins in a complex form migrated below the diagonal ‘monomer zone’ after separation by 2D-PAGE, which was also the case in the present study (Supplementary Figures S3A, S5B and S5E), p600 LRRK2 is expected to locate below the ‘monomer zone’ on 2D-PAGE gels if it represents the dimeric form of LRRK2. Thirdly, artificially degraded LRRK2 (p300 LRRK2) observed under harsh denaturing conditions (2% SDS, heating to 110°C for 30 min at pH 6.8) migrated at the ~300 kDa position on a BN-PAGE gel (Figure 3 and Supplementary Figure S4). The fact that a truncated form of LRRK2 lacking the C-terminus migrates at the position corresponding to the molecular mass of full-length LRRK2 by BN-PAGE strongly indicates that the estimation of the molecular mass of LRRK2 using standard proteins for BN-PAGE is inaccurate. Finally, p600 LRRK2 was fractionated into the ~272 kDa fraction by separation on a glycerol gradient (Figure 4). This result is also consistent with our view that p600 LRRK2 represents the monomeric form of LRRK2. We suspect that p300 LRRK2 has been mistakenly regarded as the monomeric form of LRRK2 in the previous literature [18–21], simply because its migrating position on BN-PAGE gels apparently coincided with that of full-length LRRK2.
To further investigate why the monomeric form of LRRK2 migrates at the ~600 kDa position on BN-PAGE gels, we examined other HMM proteins by BN-PAGE and found that the monomeric forms of ATM or mTOR migrate at higher positions compared with their theoretical masses (Supplementary Figures S5A and S5C). Recently, the three-dimensional structure of fully assembled mTORC1 was determined using cryo-electron microscopy . Although two molecules of mTOR were observed in one mTORC1, direct interaction between the two mTORs was not observed. Therefore p600 mTOR could not be a residual homodimer of mTOR after denaturation, suggesting that p600 mTOR represents the monomeric form of mTOR. Taken together, we hypothesized that the migrating positions of the monomeric form of HMM proteins deviate from their theoretical mass on BN-PAGE gels. Consistent with this hypothesis, 2D-PAGE analysis revealed that, at least within the HMM range (>150 kDa), the molecular mass of the proteins within the ‘monomer zone’ estimated by using standards for BN-PAGE generally deviate from that estimated using standards for SDS/PAGE, the latter being comparable with the theoretical mass of proteins, although both are relatively consistent within the low molecular mass range (Supplementary Figure S3C). In addition, there are several reports showing that the apparent molecular mass of monomeric forms of HMM proteins on BN-PAGE gels is inconsistent with their theoretical mass, which are in support of our hypothesis [31,35–37]. An in vitro kinase assay following separation on a glycerol gradient revealed that p600 LRRK2 possesses kinase activity (Figure 5B). We also confirmed that p600 LRRK2 was pulled down by GTP–agarose (Figure 5C). These results indicated that the monomeric form of LRRK2 harbours the kinase and GTP-binding activities. As reported previously , the LRRK2 K1906M mutant behaves similarly to wild-type LRRK2 on BN-PAGE gels, whereas the T1348N mutant lacking the GTP-binding activity migrates solely as the HMM substances (input of Figure 5C), suggesting that the kinase activity is dispensable for the formation of HMM substances. Although the LRRK2-positive HMM substances observed on BN-PAGE gels has not been fully characterized, the T1348N mutant might adopt an extended folding due to the lack of GTP binding, which may result in an apparently slow migration on native gels. Alternatively, it remains possible that the HMM substances could represent dimeric or oligomeric forms of LRRK2, and that the slower migration of the T1348N mutant might reflect the dependence of LRRK2 oligomerization on its GTP-/GDP-binding state, as suggested by Sen et al. .
To examine whether dimerization causes further up-regulation of the kinase activity, we employed a chemical-induced dimerization system (Figure 6A). This method is useful to examine the functional significance of dimerization of proteins, and has been utilized to show the essential role of dimerization in the activation of kinases . However, the kinase activity of LRRK2 measured by an in vitro kinase assay was not up-regulated upon treatment with AP20187 (Figure 6B). Recently, Berger et al.  reported that dimeric LRRK2 is enriched at the cell membrane, and that membrane-associated LRRK2 exhibits a greater activity compared with cytosolic LRRK2. On the basis of these findings, they hypothesized that LRRK2 forms a dimer upon translocation to the membrane and gets activated, although the causal relationships remained unclear. Our findings from the present study showed that induced dimerization of LRRK2 failed to change its intrinsic kinase activity as well as the subcellular localization (Figure 6 and Supplementary Figure S8), suggesting that dimerization itself is not sufficient for the translocation or the activation of LRRK2. The causal relationship between membrane translocation, dimerization and activation of LRRK2 should be examined further, e.g. using a chemical-induced membrane translocation experiment  in combination with BN-PAGE and an in vitro kinase assay.
In summary, we have re-examined the biochemical properties of dimerization of LRRK2 and found that LRRK2 predominantly exists as a monomer within mammalian cells. We also found that the monomeric form of LRRK2 harbours the kinase and GTP-binding activities and migrates at ~600 kDa on BN-PAGE gels. These findings suggest that a careful examination is required to investigate dimerization or complex formation of LRRK2 using BN-PAGE, since misestimation of the molecular mass of monomeric HMM proteins is unavoidable in BN-PAGE. Given that LRRK2 predominantly exists as a monomer in resting cells, it would be possible that LRRK2 dimerizes under certain conditions (e.g. extracellular stimuli, oxidative stress) or within a specific subcellular compartment. Further investigation on the dimerization of LRRK2 would provide a clue to the elucidation of the regulatory mechanism of LRRK2.
ataxia telangiectasia mutated
FKBP (FK506-binding protein) variant
familial Parkinson's disease
human embryonic kidney
higher molecular mass
leucine-rich repeat kinase 2
mammalian target of rapamycin
regulatory associated protein of mTOR
small interference RNA
Genta Ito and Takeshi Iwatsubo planned the study, analysed the results and wrote the paper. Genta Ito performed the experiments.
We thank Dr John Anderson, Dr Zhao Ren, Dr Dale Schenk and other scientists at Elan Pharmaceuticals, and current and former laboratory members for helpful discussions and technical assistance.
This work was supported by the Japan Society for the Promotion of Science [grant number 10014424] [Grant-in-Aid for Young Scientists (B)], by the Ministry of Education, Culture, Sports, Science and Technology [grant number 17025009] (Grant-in-Aid for Scientific Research on Priority Areas-Research on Pathomechanism of Brain Disorders, Global Center of Excellence Program), by the Japan Science and Technology Agency [Core Research for Evolutional Science and Technology (CREST)] and by Elan Pharmaceuticals (Innovation Program).