The G2019S pathogenic mutation disrupts sensitivity of leucine-rich repeat kinase 2 to manganese kinase inhibition

Authors


Address correspondence and reprint requests to Dr Benoit I. Giasson, Department of Pharmacology, University of Pennsylvania School of Medicine, 125 John Morgan Building, 3620 Hamilton Walk, Philadelphia, PA 19104-6084, USA. E-mail: giassonb@mail.med.upenn.edu

Abstract

J. Neurochem. (2010) 115, 36–46.

Abstract

Mutations in leucine-rich repeat kinase-2 (LRRK2) are the most common cause of late-onset Parkinson disease. Previously, we showed that the G2019S pathogenic mutation can cause a dramatic increase (approximately 10-fold) in kinase activity, far above other published studies. A notable experimental difference was the use of Mn-ATP as a substrate. Therefore, the effects of metal cation-ATP cofactors on LRRK2 kinase activity were investigated. It is shown, using several divalent metal cations, that only Mg2+ or Mn2+ can support LRRK2 kinase activity. However, for wild-type, I2020T, and R1441C LRRK2, Mn2+ was significantly less effective at supporting kinase activity. In sharp contrast, both Mn2+ and Mg2+ were effective at supporting the activity of G2019S LRRK2. These divergent effects associated with divalent cation usage and the G2019S mutation were predominantly because of differences in catalytic rates. However, LRRK2 was shown to have much lower (approximately 40-fold) ATP Km for Mn-ATP compared with Mg-ATP. Consequently, sub-stoichiometric concentrations of Mn2+ can act to inhibit the kinase activity of wild-type, but not G2019S LRRK2 in the presence of Mg2+. From these findings, a new model is proposed for a possible function of LRRK2 and the consequence of the G2019S LRRK2 pathogenic mutation.

Abbreviations used:
GST

glutathione-S-transferase

LRRK2

leucine-rich repeat kinase-2

MBP

myelin basic protein

PD

Parkinson disease

SDS

sodium dodecyl sulfate

WT

wild-type

Parkinson disease (PD) is the second most prevalent neurodegenerative disease in the developing world, and is believed to result from complex genetic and environmental factors. It is characterized clinically by resting tremor, bradykinesia, postural instability, and muscle rigidity (Hoehn and Yahr 1967; Simuni and Hurtig 2000). Although primarily viewed as an idiopathic disease, a number of gene defects have been established to either cause or increase the risk of PD (reviewed in Forman et al. 2005; Ross and Farrer 2005; Bonifati 2007; Thomas and Beal 2007; Biskup et al. 2008; Henchcliffe and Beal 2008). More specifically, autosomal dominant missense mutations in the gene for leucine-rich repeat kinase 2 (LRRK2/PARK8) are the most common known cause of PD (Paisan-Ruiz et al. 2004; Zimprich et al. 2004; Giasson and Van Deerlin 2008). At least 5 LRRK2 missense mutations are considered definitely pathogenic, but many other amino acid substitutions within this protein have been linked to PD (see Fig. 1a) (Paisan-Ruiz et al. 2004; Zimprich et al. 2004; Mata et al. 2006; Taylor et al. 2006; Biskup and West 2008; Giasson and Van Deerlin 2008). The most prevalent mutation, G2019S, is reportedly responsible for 0.6–1.6% of sporadic PD (Deng et al. 2005; Gilks et al. 2005; Kachergus et al. 2005) and 2–8% of familial PD cases (Deng et al. 2005; Di Fonzo et al. 2005; Hernandez et al. 2005; Kachergus et al. 2005; Nichols et al. 2005; Paisan-Ruiz et al. 2005). In certain ethnicities, such as North African Arabs and Ashkenazi Jews, the G2019S mutation is even a greater factor contributing to PD (22–41% of individuals with disease; Lesage et al. 2005, 2006; Ozelius et al. 2006).

Figure 1.

 Characterization of recombinant leucine-rich repeat kinase 2 (LRRK2) kinase activity. (a) Schematic of LRRK2 showing the major domains [ankyrin-like (ANK), Leu-rich repeat (LRR), Ras-in-complex (ROC), C-terminal of Ras (COR)] and the position of the mutations that are considered definitely pathogenic. (b) Western blot with anti-GST antibody showing equal amounts of glutathione affinity-purified recombinant wild-type (WT) and mutant (G2019S, I2020T and R1441C) GST-LRRK2 full-length proteins. (c) Relative kinase activity of WT, R1441C, G2019S, and I2020T LRRK2 using 200 μM ATP, 400 μM LRRKtide, and several individual divalent cations (Mg2+, Mn2+, Cd2+, Ca2+, Ni2+, Zn2+) at 5 mM. The data was standardized so that the phosphorylation reaction of LRRKtide with Mg2+ for each LRRK2 variant was normalized to 100%. (d) Comparative assessment of the ability of WT and G2019S LRRK2 to phosphorylate LRRKtide, LRRKtide-TA, or LRRKtide-YF (300 μM each) in the presence of 200 μM ATP and either 5 mM Mg2+ or Mn2+. (e) Assay demonstrating that the time-course of LRRK2 kinase activity was linear over 120 min using 200 μM ATP, 400 μM LRRKtide, and either 5 mM Mg2+ or Mn2+. For each LRRK2 variant, the activity was standardized as 100% for kinase reactions in 5 mM Mg2+ at 60 min. The error bars represent standard error of the mean.

LRRK2 is a widely expressed 2527 amino acid protein with several discrete domains (Fig. 1a) (Paisan-Ruiz et al. 2004; Zimprich et al. 2004; West et al. 2007). Containing a Ras-of-complex/GTPase domain followed by a C-terminal of Ras domain, it is a member of the ROCO protein family (see Fig. 1a). The LRRK2 kinase domain displays highest sequence homology to the mixed-linage kinase subfamily of mitogen-activated protein kinase kinase kinases, so named because of kinase sub-domain structures resembling both protein Y- and S/T-kinases (Manning et al. 2002; West et al. 2005, 2007). To date, it has been shown that LRRK2 can function as a S/T-kinase that can undergo autophosphorylation (West et al. 2005, 2007; Smith et al. 2006; Jaleel et al. 2007; Luzon-Toro et al. 2007; Greggio et al. 2008; Anand et al. 2009; Covy and Giasson 2009); although its ability to function as a Y-kinase has not been rigorously investigated. Some modeling studies have suggested that LRRK2 may be a dual specificity kinase, phosphorylating both S/T and Y residues (Manning et al. 2002; West et al. 2007), but so far it has been shown to function predominantly as a S/T-kinase (West et al. 2007; Anand et al. 2009) and only weak activity toward the Y-kinase substrate poly(E)tyrosine was reported (West et al. 2005). Furthermore, the biological functions and regulation of LRRK2, and the effects of disease-causing mutations therein are still ill-defined (Biskup and West 2008; Greggio and Cookson 2009; Webber and West 2009). For example, the R1441C mutation was demonstrated to increase kinase activity in some studies (West et al. 2005, 2007), but others have reported no significant change (Greggio et al. 2006; Jaleel et al. 2007; Anand et al. 2009; Gloeckner et al. 2009). The I2020T mutation was documented to either modestly increase (Gloeckner et al. 2006, 2009; West et al. 2007), show no change (Anand et al. 2009) or decrease kinase activity (Jaleel et al. 2007). Most studies of the G2019S mutation demonstrated increased kinase activity, although modest (two- to threefold) (West et al. 2005, 2007; Greggio et al. 2006; Jaleel et al. 2007; Anand et al. 2009; Gloeckner et al. 2009). Recently, we have shown that in one experimental paradigm, the G2019S LRRK2 mutant can demonstrate 10-fold greater kinase activity than wild-type (WT) LRRK2 (Covy and Giasson 2009). One notable difference is that we used Mn2+ as an ATP cofactor, while most other published studies have used Mg2+. Therefore, in this study, we assessed the relative kinetic effects of Mg2+ versus Mn2+ on the catalytic properties of WT LRRK2 and some disease-causing mutants thereof.

Materials and methods

Materials

Goat anti-glutathione-S-transferase (GST) polyclonal antibody was purchased from Amersham Biosciences (Piscataway, NJ, USA). The shuttling vector pCR8/GW/TOPO and the mammalian expression GST-tagged vector pDEST27 were purchased from Invitrogen (Carlsbad, CA, USA). LRRKtide (RLGRDKYKTLRQIRQ), LRRKtide-TA (RLGRDKYKALRQIRQ) that is deficient in S/T residues, LRRKtide-YF (RLGRDKFKTLRQIRQ) that is deficient in Y residues, and Nictide (RLGWWRFYTLRRARQGNTKQR) were synthesized and purified on reverse phase HPLC by GenScript USA Inc. (Piscataway, NJ, USA). The synthetic peptides myelin basic protein (MBP) fragment 104–118 (GKGRGLSLSRFSWGA), Kemptide (LRRASLG), and caesin kinase 2 substrate (RRRADDSD) were purchased from Sigma-Aldrich (St. Louis, MO, USA). The synthetic peptides MBP fragment 4–14 (KRPSQRSKYL), MBP fragment 94–102 (APRTPGGRR), MARCKS PSD-derived peptide (KKRFSFKKSFKL), and protein kinase Cδ peptide substrate (RFAVRDMRQTVAVGVIKAVDKK) were obtained from Calbiochem/EMD Biosciences (Gibbstown, NJ, USA). Kinase inhibitors JAK3 VI, K252-A, PKR, ROCK-IV, and staurosporine were purchased from Calbiochem/EMD Biosciences, and sunitinib was purchased from Sigma-Aldrich.

Cell culture

Human embryonic kidney 293T cells were cultured in Dulbecco’s modified medium with high glucose (4.5 g/L) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 U/mL streptomycin, and 2 mM l-glutamine.

LRRK2 expression constructs

The full-length human LRRK2 cDNA was amplified by PCR using Taq polymerase AccuPrime SuperMix (Invitrogen) and cloned by topoisomerase reaction into the shuttling vector pCR8/GW/TOPO. To generate the cDNAs encoding the G2019S or I2020T mutations, the LRRK2 cDNA fragment spanning the AvrII and NcoI restriction sites in LRRK2 was amplified by PCR and cloned by topoisomerase reaction into the vector pCR4-TOPO (Invitrogen). To generate the cDNA encoding the R1441C mutation, the LRRK2 cDNA fragment spanning the NdeI and SpeI restriction sites in LRRK2 was amplified by PCR and cloned by topoisomerase reaction into the vector pCR4-TOPO (Invitrogen). The mutations corresponding to the R1441C, G2019S, or I2020T amino acid substitutions, respectively, were generated using the QuickChange® Site Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). The AvrII/NcoI DNA fragments containing either the G2019S or I2020T mutants or the NdeI/SpeI DNA fragments containing the R1441C mutation were reintroduced into full-length LRRK2 by subcloning with these restriction enzymes. The sequence of the plasmids was verified by DNA sequencing using primers that span the whole cDNA as a service offered by the DNA Sequencing Facility of the University of Pennsylvania. WT and mutant full-length LRRK2 cDNAs were introduced into the pDEST27 vector by recombinase reaction using LR Clonase II enzyme (Invitrogen) to generate a plasmid expressing N-terminal GST-tagged protein.

Western blotting analysis

Proteins were resolved on sodium dodecyl sulphate (SDS)-polyacrylamide gels by SDS–polyacrylamide gel electrophoresis and electrophoretically transferred onto nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA, USA) in buffer containing 190 mM glycine, 25 mM Tris-base, and 10% methanol. Membranes were blocked with a 5% powdered skimmed milk solution dissolved in Tris-buffered saline (50 mM Tris, pH 7.6, 150 mM NaCl), incubated with anti-GST antibody followed with an anti-goat antibody conjugated to horse radish peroxidase, developed with Western Lightning Chemiluminescence Reagents (PerkinElmer Life Sciences (Boston, MA, USA), and exposed onto X-Omat Blue XB-1 films (Kodak, Rochester, NY, USA).

In vitro LRRK2 kinase assays

Human embryonic kidney 293T cells were transiently transfected with pDEST27/LRRK2 expression plasmid using calcium phosphate precipitation method buffered with N,N-bis(2-hydroxyethyl)-2-amino-ethanesulfonic acid (Chen and Okayama 1997). Cells were washed and harvested with phosphate-buffered saline 48–72 h after transfection, and re-suspended in lysis buffer (25 mM Tris pH 7.4, 150 mM NaCl, 5 mM EDTA, 10 mM beta-glycerol phosphate, 1 mM NaVO4, 1% Triton X-100, 0.5% glycerol with protease inhibitor cocktail) at 4°C. Cell debris were removed by sedimentation at 13 000 g for 15 min, and supernatants were pre-cleared by incubation with sepharose beads that were removed by sedimentation. Supernatants were incubated with glutathione-sepharose beads (GE Healthcare) for 3 h at 4°C. Beads were extensively washed with lysis buffer (five times) and wash buffer (25 mM Hepes, pH. 7.4, 1 mM dithiothreitol, 10 mM β-glycerophosphate; five times) and eluted with wash buffer with 20 mM glutathione. The kinase reactions were conducted at 25°C by incubating purified GST-LRRK2 in 25 μL of kinase buffer (25 mM Hepes, pH. 7.4, 10 mM β-glycerophosphate), with the added specified divalent cation, and varying [γ-32P]ATP (4 Ci/mmol) (1 μM to 800 μM) and LRRKtide peptide (1 μM to 1200 μM) concentrations as specified for each experiment. All reactions were conducted for 60 min at 25°C, except for the time-course experiments that were conducted for the indicated times. Reactions were applied to individual 2.5 cm-diameter disks of P-81 phosphocellulose filter paper (Schleicher & Schuell, Keene, NH, USA) that were immediately immersed in 75 mM phosphoric acid. After extensive wash with 75 mM phosphoric acid, P-81 filters were rinsed with acetone and allowed to air dry. Filters were immersed in Cytoscint liquid scintillation cocktail (Fisher Scientific, Pittsburgh, PA, USA) and 32P radioactivity on each filter was measured by liquid scintillation using an LS6500 counter (Beckman Coulter, Brea, Ca, USA). A unit of LRRK2 activity was defined as the amount of enzyme that catalyzed the incorporation of 1 pmol of 32P into LRRKtide. Km and Vmax parameters were calculated using Graph-Pad Prism v5.01 (GraphPad Software, San Diego, CA, USA).

For autophosphorylation of LRRK2, reactions were stopped with the addition of SDS sample buffer and heating to 100°C for 5 min. Samples were resolved onto 6% SDS–polyacrylamide gels, dehydrated in a 50% methanol, 10% glycerol, 10% acetic acid solution, dried, and exposed to a PhosphorImager plate (Molecular Dynamics, Piscataway, NJ, USA), and visualized using ImageQuant software (Molecular Dynamics, Inc., Sunnyvale, CA, USA).

Results

We recently generated a novel system to purify substantial amounts of active recombinant full-length LRRK2 expressed in 293T mammalian cells (Covy and Giasson 2009). In this system, LRRK2 is expressed as an N-terminal-tagged GST protein that is purified using glutathione-sepharose beads as described in ‘Materials and methods’. The kinase activity of this enzyme was previously characterized for specificity using a kinase dead version as well as kinase inhibitors in the presence of Mn2+ (Covy and Giasson 2009). The specificity of this activity was further confirmed using a series of previously published and characterized LRRK2 kinase inhibitors (Anand et al. 2009; Nichols et al. 2009; Reichling and Riddle 2009) in the presence of Mg2+ (Fig. S1). For the studies described here, WT and three pathogenic mutants of LRRK2 (R1441C, G2019S, and I2020T) were purified to equivalent levels (Fig. 1b). The kinase reactions in the present studies were conducted using the synthetic peptide LRRKtide that spans residue T558 of moesin, which was previously identified as a robust in vitro substrate for LRRK2 (Jaleel et al. 2007; Anand et al. 2009; Covy and Giasson 2009).

Protein kinases require the formation of an ATP-divalent metal complex for the phosphoryl transfer of the γ-phosphate of ATP to a protein substrate (Knowles 1980). Typically, Mg2+ serves as the essential metal ion for catalysis, however, Mn2+ and other divalent cations may support nucleotide binding and subsequent use as a cofactor for phosphoryl transfer (Courtneidge 1985; Toru-Delbauffe et al. 1986; Koland and Cerione 1990; Elberg et al. 1995; Sun and Budde 1997). Several divalent metal cations were tested to assess their ability to act as an ATP cofactor for WT and pathogenic variants of LRRK2. This analysis was conducted for pathological mutations within the active site (G2019S and I2020T) as well as a pathological mutation distinct from this site (R1441C). Of the divalent metal cations assessed, Mg2+ demonstrated the greatest ability to support LRRK2 kinase activity under the reaction conditions used here (200 μM ATP, 400 μM LRRKtide; Fig. 1c). Mn2+ also demonstrated a modest ability to promote LRRK2 kinase activity of WT, R1441C, and I2020T LRRK2, however, this property was greatly enhanced for the G2019S LRRK2 mutant. The other divalent cations tested were not able to support LRRK2 kinase activity.

It is well known that many Y-kinases prefer Mn2+ as an ATP cofactor (Courtneidge 1985; Hunter and Cooper 1985; Koland and Cerione 1990; Elberg et al. 1995; Sun and Budde 1997), and for some dual specificity kinases, the presence of Mn2+ versus Mg2+ can cause a switch in preferential activity from a protein S/T-kinase to a protein Y-kinase (Yuan et al. 1993; Huang et al. 1994). Therefore, it is possible that Mn2+ may support the activity of LRRK2 as a Y-kinase since LRRKtide (RLGRDKYKTLRQIRQ) contains both a T and a Y residue. To further investigate the type of kinase activities associated with the use of Mg2+ or Mn2+, kinase reactions for WT, R1441C, G2019S, and I2020T LRRK2 were performed in the presence of 5 mM Mg2+ or Mn2+ and LRRKtide, LRRKtide deficient in S/T residues (LRRKtide-TA; RLGRDKYKALRQIRQ) or LRRKtide that is deficient in Y residues (LRRKtide-YF; RLGRDKFKTLRQIRQ). Results showed that only LRRKtide with a T residue served as a substrate for LRRK2 in the presence of Mg2+ or Mn2+ (Fig. 1d). To further confirm that the persistence of kinetic activity of the G2019S mutant in the presence of Mn2+ is not a result of altered substrate specificity, a number of other characterized synthetic peptides with known kinase recognition motifs, including the engineered LRRK2 substrate Nictide (Nichols et al. 2009), were analyzed in the presence of Mg2+ or Mn2+, and no differences in substrate recognition were observed (Fig. S2).

Time-course analysis of the kinase activity for WT, R1441C, G2019S, and I2020T LRRK2 in the presence of Mg2+ or Mn2+ was performed to determine the consistent progress of the phosphoryl transfer over time. Results showed that the kinase reaction was linear over the course of 2 h, allowing extrapolation of initial velocity and confirming that differences in kinetic activity were not because of changes in kinase stability (Fig. 1e). In addition, similar to our initial screen of divalent cations, WT, R1441C, G2019S, and I2020T LRRK2, kinase activity was much greater in the presence of Mg2+ compared with Mn2+ for all enzymes except the G2019S mutant.

To further investigate the effects of Mg2+ and Mn2+ on LRRK2 kinase activity, the ability and extent of each ion to support kinase activity was assessed by varying their concentrations. For all variants of LRRK2 assayed against LRRKtide, increasing the concentration of Mg2+ resulted in increased kinase activity with maximum activity at 5 mM. Mn2+ also supported the kinase activity of LRRK2, but maximal activity was observed at ∼ 0.5 mM and all variants demonstrated a similar relative decrease in activity at higher concentrations with approximately twofold lower activity at 5 mM Mn2+ compared with 0.5 mM Mn2+ (Fig. 2). R1441C and I2020T mutants demonstrated Mn2+ concentration activity profiles that were similar to WT LRRK2, but G2019S LRRK2 revealed dramatically greater activity than the other LRRK2 variants in the presence of Mn2+ and this was especially striking in the presence of Mn2+ below or near sub-stoichiometric concentrations of ATP. These effects were replicated using the synthetic substrate Nictide (Fig. S3). Again, Mn2+ is capable of driving kinase activity, although not to the extent of Mg2+ for WT, R1441C, and I2020T LRRK2. For G2019S LRRK2, Mn2+ demonstrated the ability to drive much greater kinase activity than for the other LRRK2 variants, and in fact, G2019S LRRK2 showed greater kinase activity in the presence of Mn2+ compared with Mg2+ (Fig. S3C). Although the effects of Mg2+ and Mn2+ on WT, R1441C, G2019S, and I2020T LRRK2 kinase activity using either LRRKtide or Nictide are consistent, there are some relative differences and more detailed kinetic studies of the difference between LRRKtide and Nictide are under investigation.

Figure 2.

 Concentration dependent affects of Mg2+ and Mn2+ on the kinase activity of WT, R1441C, G2019S, and I2020T leucine-rich repeat kinase 2 (LRRK2) on LRRKtide phosphorylation. Kinase reactions for (a) WT, (b) R1441C, (c) G2019S and (d) I2020T LRRK2 were conducted using 200 μM ATP, 400 μM LRRKtide and varying concentrations of Mg2+ or Mn2+ between 0.1 mM and 10 mM. The error bars represent standard error of the mean.

Similar results for divalent cation preference were observed for autophosphorylation of LRRK2. Increasing concentrations of Mg2+ drove robust autophosphorylation of both WT and G2019S LRRK2, and these levels were greatly reduced in the presence of Mn2+ for WT, but not G2019S (Fig. 3a). Interestingly, 0.1 mM Mn2+ induced higher autophosphorylation activity in both WT and G2019S LRRK2 compared with Mg2. This selectivity may reflect higher affinity for ATP-Mn (see below), but direct affects of divalent cations on LRRK2 autophosphorylation cannot be excluded and this will require further detailed analyses. Nevertheless, consistent with LRRKtide, WT LRRK2 exhibited lower autophosphorylation activity in 5 mM Mn2+ versus 5 mM Mg2+, while G2019S LRRK2 displayed similar autophosphorylation in the presence of either 5 mM Mn2+ or Mg2+ (Fig. 3b).

Figure 3.

 Concentration dependent affects of Mg2+ and Mn2+ on the autophosphorylation activity of wild-type (WT) and G2019S leucine-rich repeat kinase 2 (LRRK2). (a) Kinase reactions for WT and G2019S were conducted using 100 μM ATP and with varying concentrations of divalent cation between 0.1 and 10 mM. Autoradiography shows the effects of increased concentration of Mg2+ or Mn2+ autophosphorylation activity. (b) The relative activity of WT and G2019S LRRK2 when using either Mg2+ or Mn2+ as cofactor at 5 mM. The error bars represent standard error of the mean.

To investigate the nature of the differences in the kinetics properties affected by Mg2+ versus Mn2+ and the provocative effects of the G2019S LRRK2 pathogenic mutation, the Michaelis–Menton parameters were determined for WT, R1441C, I2020T, and G2019S LRRK2 using LRRKtide as a substrate. To ensure that the ATP-cation complex was at saturating levels, 5 mM Mg2+, or Mn2+ was used to remain at a concentration several fold above the highest concentration of ATP. Furthermore, results showed at this concentration of Mg2+, maximal LRRK2 activity is obtained. Kinetic analyses for the Km of LRRKtide, while keeping the ATP concentration constant at 200 μM, revealed values consistent with previous reports utilizing Mg2+ and N-terminally truncated versions (LRRK21326-2527 or LRRK2970-2527) of LRRK2 (Jaleel et al. 2007; Anand et al. 2009; Nichols et al. 2009). The affinity for LRRKtide was not dramatically affected by the use of Mg2+ and Mn2+, or by mutations in LRRK2; although some subtle changes were observed (Fig. 4 and Table 1). The apparent Vmaxs measured under these conditions revealed that WT, I2020T, and R1441C LRRK2 had much greater relative catalytic rates (approximately 10-fold) in the present of Mg2+ versus Mn2+. In regards to each respective cation, the R1441C and I2020T mutants resulted in only small changes in catalytic rate compared with WT, while the apparent Vmax for G2019S in the presence of Mg2+ was approximately 2.5-fold greater than WT LRRK2. However, the most dramatic effect was the relative maintenance of catalytic activity of G2019S LRRK2 in the presence of Mn2+ relative to Mg2+. This effect was unique to the G2019S mutation. In fact, for G2019S LRRK2, the Vmax in the presence of Mn2+ compared with Mg2+ was only approximately 1.5-fold lower, in contrast to the approximately ninefold lower activity for WT, R1441C, and I2020T LRRK2.

Figure 4.

 Kinetic characteristics of wild-type (WT) and leucine-rich repeat kinase 2 (LRRK2) mutations while varying the concentration of LRRKtide as a substrate. Assessment of the kinetic features of (a) WT, (b) R1441C, (c) G2019S, and (d) I2020T recombinant LRRK2 with ATP at 200 μM with either 5 mM Mg2+ or 5 mM Mn2+. The concentration of LRRKtide was varied as indicated in the x-axis. The error bars represent standard error of the mean.

Table 1.   Michaelis–Menton parameters of wildtype and leucine-rich repeat kinase 2 mutations for LRRKtide
 KmVmax
MgMnMgMn
  1. The relative apparent Vmax (U/min) and Km (μM) for LRRKtide were assessed with ATP at 200 μM and Mg2+ or Mn2+ at 5 mM. The errors represent standard deviations.

Wildtype208.9 ± 11.50242.6 ± 14.742.57 ± 0.0460.304 ± 0.005
R1441C208.6 ± 17.30300.5 ± 30.553.63 ± 0.1020.286 ± 0.009
G2019S243.8 ± 19.64235.8 ± 13.096.80 ± 0.2435.222 ± 0.084
I2020T251.7 ± 37.80255.5 ± 35.913.31 ± 0.1460.282 ± 0.012

The determination of the apparent Km of ATP while keeping LRRKtide constant at 400 μM revealed that WT LRRK2 and LRRK2 mutants have a significantly lower ATP Km in the presence of Mn2+ compared with Mg2+ (Fig. 5 and Table 2). This suggests a greater affinity for Mn-ATP than Mg-ATP. ATP Km for WT and R1441C had similar properties, while I2020T LRRK2 had a relative lower ATP Km in the presence of either Mg2+ or Mn2+, respectively. Conversely, G2019S LRRK2 demonstrated relatively higher ATP Km in the presence of either Mg2+ or Mn2+, respectively, compared with WT LRRK2. Again, the analyses to assess Vmax values showed that WT LRRK2 demonstrated a greater catalytic rate in the presence Mg2+ versus Mn2+, however, the most striking finding was the much higher Vmax for G2019S LRRK2 in the presence of Mn2+ compared with WT and other mutants of LRRK2.

Figure 5.

 Kinetic characteristics of wild-type (WT) and leucine-rich repeat kinase 2 (LRRK2) mutations while varying the concentration of ATP as a substrate. Assessment of the kinetic features of (a) WT, (b) R1441C, (c) G2019S, and (d) I2020T recombinant LRRK2 with LRRKtide at 400 μM with either 5 mM Mg2+ or 5 mM Mn2+. The concentration of ATP was varied as indicated in the x-axis. The error bars represent standard error of the mean.

Table 2.   Michaelis–Menton parameters of wildtype and leucine-rich repeat kinase 2 mutations for ATP. The relative apparent Vmax (U/min) and Km (μM) for ATP were assessed with LRRKtide at 400 μM and Mg2+or Mn2+at 5 mM. Errors represent standard deviations
 KmVmax
MgMnMgMn
Wildtype54.53 ± 4.731.40 ± 0.1732.08 ± 0.0440.189 ± 0.003
R1441C62.97 ± 4.022.61 ± 0.4723.06 ± 0.0490.238 ± 0.006
G2019S74.08 ± 7.733.93 ± 0.4364.68 ± 0.1263.315 ± 0.057
I2020T18.13 ± 2.630.95 ± 0.1532.67 ± 0.0780.179 ± 0.004

Since under physiological conditions, Mn2+ exists at sub-stoichiometric concentrations relative to Mg2+, we compared the relative effects of varying sub-stoichiometric concentrations of Mn2+ in the presence of a fixed concentration of Mg2+. For these experiments, Mg2+ was maintained at 5 mM while the concentration of Mn2+ was varied between 0 and 5 mM. Increasing the concentration of Mn2+ resulted in reduced kinase activity for WT, R1441C, and I2020T LRRK2 (Fig. 6a, b and d), consistent with a greater affinity for Mn-ATP versus Mg-ATP paired with a reduction in catalytic activity when utilizing Mn-ATP. In sharp contrast, increases in the concentration of Mn2+ between 0.1 and 1 mM increased the kinase activity of G2019S LRRK2, while only modestly reducing kinase activity at 5 mM Mn2+ (Fig. 6c). Given that G2019S LRRK2 levels of kinase activity in the presence of Mn2+ reaches a maximum at ∼1 mM and that this activity is similar to G2019S LRRK2 in the presence of 5 mM Mg2+ (Fig. 2c), but that 5 mM Mg2+ in the presence of 0.1 or 0.5 mM Mn2+ results in higher activity, it is possible that Mn2+ may additionally modulate kinase activity by acting at a second site. Overall, these results are consistent with those reported recently (Lovitt et al. 2010) and show that at millimolar concentrations of Mg2+, micromolar concentrations of Mn2+ decrease the levels of LRRK2 kinase activity, while the G2019S mutation causes a loss of this negative regulation, if not reversing the effect and increasing activity.

Figure 6.

 Analysis of effects of sub-stoichiometric concentrations of Mn2+ on wild-type (WT) and mutant leucine-rich repeat kinase 2 (LRRK2) kinase activity driven by Mg2+. The activity of (a) WT, (b) R1441C, (c) G2019S, and (d) I2020T recombinant LRRK2 was assessed in the presence of 200 μM ATP, 400 μM LRRKtide, and 5 mM Mg2+ with concentrations of Mn2+ varying from 0 to 5 mM. As a reference, reactions with only 5 mM Mn2+ are also shown. For each LRRK2 variant, the activity was standardized as 100% activity for kinase reactions with only 5 mM Mg2+. The error bars represent standard error of the mean.

Discussion

In stark contrast to several other published findings (West et al. 2005, 2007; Greggio et al. 2006; Jaleel et al. 2007; Anand et al. 2009; Gloeckner et al. 2009), we recently demonstrated that the G2019S mutation in LRRK2 could result in a much greater kinase activity relative to WT LRRK2 (Covy and Giasson 2009). A notable difference in our study was the use of Mn2+ as an ATP cofactor compared with the use of Mg2+ in other studies. This led us to investigate the ability of divalent metal cations to support the kinase activity of WT and several pathogenic mutants of LRRK2. It was shown that only Mg2+ or Mn2+ could be used as ATP cofactors to support the kinase activity of LRRK2, and for WT, I2020T, and R1441C LRRK2, Mg2+ was a much better cofactor at promoting kinase activity. Remarkably, both Mn2+ and Mg2+ were effective at promoting the activity of G2019S LRRK2. Using LRRKtide peptides deficient in either T or Y resides, it was shown that WT and mutants of LRRK2 appear to be exclusively S/T-protein kinases, and Mn2+ does not change this specificity. Furthermore, no changes in specificity were observed for a number of known generic kinase substrates for WT or G2019S LRRK2 when utilizing Mn2+ or Mg2+. Interestingly, LRRK2 demonstrated greater activity toward the LRRKtide-YF peptide compared with LRRKtide, however, this is consistent with a previous study mapping the preferred consensus motif of LRRK2 substrates where it has been shown that for the P-2 position, an F is favored over a Y residue (Nichols et al. 2009).

Assessment of Michaelis–Menton parameters showed that the use of Mn2+ versus Mg2+ did not dramatically affect LRRKtide Km. Consistent with previous reports using N-terminally truncated versions of LRRK2 (Anand et al. 2009; Reichling and Riddle 2009), ATP Km for G2019S LRRK2 was higher compared with WT LRRK2, while that of I2020T LRRK2 was lower in the presence of Mg2+. WT and all mutants of LRRK2 demonstrated significantly lower ATP Km for Mn-ATP compared with Mg-ATP, and similar findings have been attributed predominantly to a high affinity for Mn-ATP (Lovitt et al. 2010). WT, R1441C, and I2020T LRRK2 showed slower catalytic rates in the presence of Mn2+ compared with Mg2+. A dramatically higher Vmax for G2019S in the presence of Mn2+ compared with WT LRRK2 is the predominant reason for the approximately 10-fold greater level of kinase activity for G2019S LRRK2 versus WT LRRK2 that we previously reported (Covy and Giasson 2009). Consistent with these data, similar findings were observed in a recent study using a truncated version (LRRK2970-2527) of LRRK2 expressed and purified from baculovirus-infected insect cells (Lovitt et al. 2010). However, the differences in ATP Km for the different LRRK2 variants also indicate that relative kinase activity can be affected depending on the concentration of ATP used in the reactions, especially around or below the Km. In the presence of Mg2+, the relative difference in kinase activity for G2019S versus WT LRRK2 is usually approximately twofold greater as reflected by the Vmax values.

The molecular basis for the differences in catalytic rates between using Mg2+ versus Mn2+ are not clear and will require comprehensive structural analysis. However, it is possible that both metal cations may coordinate the placement and interactions of divalent cation-ATP in the binding site differently, as such affecting the rate of the phosphoryl-transfer step. Alternatively, reduced rate of cation-ADP release, which can be a rate limiting step in protein kinase phosphorylation, may be the major property affected by Mn2+. Indeed, Lovitt et al. (2010) have shown that LRRK2 displays a higher affinity for ADP in the presence of Mn2+ compared with Mg2+.

The dramatic alterations in the catalytic properties of LRRK2 in the presence Mn2+ caused by the G2019S mutation and the greater activity of this mutant is conceptually consistent with the role of this residue in the active site. G2019 is part of the highly conserved DFG-motif in subdomain VII of the kinase domain (Hanks et al. 1988; Kannan and Neuwald 2005), and the D residue in this motif is involved in Mg2+ binding and proper coordination of Mg-ATP in the active site (De Bondt et al. 1993; Karlsson et al. 1993; Taylor and Radzio-Andzelm 1994; Levinson et al. 2006). The relative higher ATP Km, especially in the presence of Mn2+, compared with WT LRRK2, suggests that a more rapid exchange of Mn-ATP/ADP at the active site may be partially responsible for the increased catalytic rate of G2019S LRRK2 in the presence of Mn2+. The DFG-motif is located at the N-terminal hinge region of the activation loop that switches from an open and extended conformation in the active state to a more closed conformation in the inactive state (De Bondt et al. 1993; Taylor and Radzio-Andzelm 1994; Kannan and Neuwald 2005), and this G residue is thought to play an important role in inducing proper orientation of the D residue (Kornev et al. 2006). After inactivation, the G residue usually performs an extreme twist, thereby facilitating this D residue to turn away from the catalytic site. The lack of a side chain is thought to allow the G residue to make this turn with little steric hindrance. The G2019S mutation may disrupt the ability of this movement, and could keep the D residue positioned in the active site for longer activation periods that may also contribute to a greater catalytic rate. Nevertheless, comprehensive structural analyses will be required to better understand the molecular affects of the G2019S mutation.

The I2020T mutation, which changes the first residue following the DFG-motif, did not have a dramatic affect on the kinase kinetic properties as did the G2019S mutation. Nevertheless, consistent with other studies (Reichling and Riddle 2009), the I2020T mutation demonstrated significantly reduced ATP Km for ATP.

Studies of the effects of Mg2+ on the observed LRRK2 kinase activity indicate that increasing the concentration of Mg2+ beyond what is needed to saturate total ATP results in further activation of LRRK2. Lovitt et al. (2010) recently reported comparable data using a truncated version of LRRK2 (LRRK2970-2527). Such an effect has been observed for several other kinases and in some cases has been attributed to a second Mg2+ binding site (Sun and Budde 1997; Saylor et al. 1998). Increasing the concentration of Mg2+ or Mn2+ up to 0.5 mM in the presence of 200 μM ATP increases LRRK2 kinase activity, but that is likely predominantly because of increasing the effective concentration of the cation-ATP complex. The data from Mg2+/Mn2+ competition assays with G2019S LRRK2 also suggest that between 0.1 and 0.5 mM, Mn2+ may also promote the kinase activity of LRRK2 independently of increasing Mn-ATP concentration. In contrast to Mg2+, increasing the concentration of Mn2+ beyond 2 mM had an inhibitory affect on LRRK2 kinase activity, and a similar inhibitory effect observed by Lovitt and colleagues was attributed to decreased catalytic turnover (Lovitt et al. 2010). In our studies, all LRRK2 variants demonstrated similar relative decreases in activity with increasing Mn2+ beyond 2 mM, although for G2019S LRRK2, kinase activity remained relatively high at all concentrations of Mn2+ compared with WT and other mutants of LRRK2.

These enzymatic kinetic properties of LRRK2 may be more than just interesting in vitro experiments and may reflect important characteristics of this enzyme that were designed for biological functionality. Manganese is an essential metal that is widely used in certain manufacturing industries. Environmental or occupational exposure to high manganese levels can cause neurotoxicity with the development of a form of parkinsonism known as manganism (Dobson et al. 2004; Perl and Olanow 2007). Mn2+ is required for many enzymes and its levels in tissue are usually stable, but it can accumulate in certain brain regions, including the basal ganglia, following elevated exposure (Dobson et al. 2004; Perl and Olanow 2007). The homeostatic level of free Mn2+ in cells is maintained by various transporters and by binding to various proteins (Dobson et al. 2004; Au et al. 2008). It is widely believed that in vivo, Mg2+ is the major ATP cofactor involved in kinase reactions because of its higher abundance compared with any other divalent metal cation. However, given the much greater affinity of LRRK2 for Mn-ATP compared with Mg-ATP (approximately 40-fold lower Km), and our studies of the effects of sub-stoichiometric concentration of Mn2+ on Mg2+-mediated LRRK2 kinase activity, it is possible that Mn2+ may act as a potent inhibitor of LRRK2 kinase activity in vivo. As seen here, in the presence of Mg2+, co-incubation with 100-fold lower concentrations of Mn2+ cause a significant reduction in the kinase activity of WT LRRK2; however, this effect does not apply to the G2019S mutant, where no activity is lost. Therefore, we hypothesize that LRRK2 may act as a sensor for increased cytoplasmic Mn2+ levels, which results in decreased kinase activity that would modulate downstream counteractive measures. As a consequence of its alterations in kinetic properties, it is predicted that the G2019S mutant would remain largely active at physiologically elevated Mn2+ levels and therefore this putative signaling pathway should be compromised. Although it can be difficult to directly demonstrate altered changes in enzyme activity in vivo as a consequence of alterations in specific ion levels, some in vivo studies have shown that elevated Mn2+ levels can decrease the activity of some kinases (Puli et al. 2006). Direct in vivo studies of the proposed model of the effects of Mn2+ on WT and G2019S LRRK2 kinase activity are complicated by the many conflicting studies of the effects of LRRK2 on signal transduction pathways and the lack of robust specific in vivo markers of LRRK2 activity (Biskup and West 2008; Greggio and Cookson 2009; Webber and West 2009), but in vivo studies of Mn2+ affects on LRRK2 kinase activity are currently under investigation. Nevertheless, the studies described here provide important insights into the kinetic properties of the kinase activity of LRRK2 and pathogenic mutants as affected by divalent metal cations.

Acknowledgements

This work was funded by grants from the National Institute of Neurological Disorders and Stroke (NS053488), the University of Pennsylvania Institute on Aging and The Ellison Medical Foundation.

Ancillary