Human Genetics

, Volume 120, Issue 6, pp 857–863

The LRRK2 Gly2385Arg variant is associated with Parkinson’s disease: genetic and functional evidence

Authors

    • Department of NeurologySingapore General Hospital
    • National Neuroscience Institute
    • SingHealth Research
  • Y. Zhao
    • Department of Clinical ResearchSingapore General Hospital
  • L. Skipper
    • Population GeneticsGenome Institute of Singapore
  • M. G. Tan
    • Department of Clinical ResearchSingapore General Hospital
  • A. Di Fonzo
    • Department of Clinical GeneticsErasmus MC Rotterdam
  • L. Sun
    • Department of Clinical ResearchSingapore General Hospital
  • S. Fook-Chong
    • Department of Clinical ResearchSingapore General Hospital
  • S. Tang
    • Knowledge Discovery DepartmentInstitute for Infocomm Research
  • E. Chua
    • Department of NeurologySingapore General Hospital
  • Y. Yuen
    • Department of Health ScreeningSingapore General Hospital
  • L. Tan
    • National Neuroscience Institute
  • R. Pavanni
    • Department of NeurologySingapore General Hospital
    • SingHealth Research
  • M. C. Wong
    • Department of NeurologySingapore General Hospital
    • SingHealth Research
  • P. Kolatkar
    • Population GeneticsGenome Institute of Singapore
  • C. S. Lu
    • Department of Neurology, Chang Gung Memorial Hospital, College of MedicineChang Gung University
  • V. Bonifati
    • Department of Clinical GeneticsErasmus MC Rotterdam
  • J. J. Liu
    • Population GeneticsGenome Institute of Singapore
Original Investigation

DOI: 10.1007/s00439-006-0268-0

Cite this article as:
Tan, E.K., Zhao, Y., Skipper, L. et al. Hum Genet (2007) 120: 857. doi:10.1007/s00439-006-0268-0

Abstract

Evidence of LRRK2 haplotypes associated with Parkinson’s disease (PD) risk was recently found in the Chinese population from Singapore, and a common LRRK2 missense variant, Gly2385Arg, was independently detected as a putative risk factor for PD in the Chinese population from Taiwan. To test the association between the Gly2385Arg variant in a large case-control sample of Chinese ethnicity from Singapore, and to perform functional studies of the wild type and Gly2385Arg LRRK2 protein in human cell lines. In a case-control study involving 989 Chinese subjects, the frequency of the heterozygous Gly2385Arg genotype was higher in PD compared to controls (7.3 vs. 3.6%, odds ratio = 2.1, 95% CI: 1.1–3.9, P = 0.014); these values yield an estimated population attributable risk (PAR) of ∼4%. In a multivariate logistic regression analysis with the disease group (PD vs. controls) as the dependent variable and the genotype as an independent factor with adjustments made for the effect of age and gender, the heterozygous Gly2385Arg genotype remained associated with an increased risk of PD compared to wild type genotype (odds ratio = 2.67, 95% CI: 1.43–4.99, P = 0.002). The glycine at position 2385 is a candidate site for N-myristoylation, and the Gly2385Arg variant replaces the hydrophobic glycine with the hydrophilic arginine, and increases the net positive charge of the LRRK2 WD40 domain. In transfection studies, we demonstrated that both the wild type and Gly2385Arg variant LRRK2 protein localize to the cytoplasm and form aggregates. However, under condition of oxidative stress, the Gly2385Arg variant was more toxic and associated with a higher rate of apoptosis. Our study lends support to the contention that the Gly2385Arg is a common risk factor for PD in the Chinese population. Our bioinformatics and in-vitro studies also suggest that the Gly2385Arg variant is biologically relevant and it might act through pro-apoptotic mechanisms.

Introduction

Parkinson’s disease (PD) is the second most common neurodegenerative disease. Mutations in the leucine-rich repeat kinase 2 (LRRK2, PARK8) are the most frequent known cause of familial autosomal dominant PD (Zimprich et al. 2004; Paisan-Ruiz et al. 2004). The common G2019S mutation accounts for 3–7% of familial PD and 1–3% in sporadic PD in several ethnic populations, with highest prevalence in North Africans and Ashkanezi Jews (Skipper et al. 2005a; Di Fonzo et al. 2005; Gilks et al. 2005; Kachergus et al. 2005; Lesage et al. 2005, 2006; Ozelius et al. 2006; Goldwurm et al. 2005). Information on the significance of common LRRK2 polymorphisms in relationship to risk of sporadic PD is currently limited to a few ethnic populations. Three studies suggest little role of these variants in Caucasian PD populations (Biskup et al. 2005; Paisan-Ruiz et al. 2005, 2006). However, we have shown that specific LRRK2 haplotypes may modulate risk of PD in the Chinese population (Skipper et al. 2005b). The absence of the common G2019S mutations in three independent Chinese populations involving >2,000 study subjects (Tan et al. 2005a; Lu et al. 2005; Fung et al. 2006) suggests the possibility that ethnicity specific differences may exist for other LRRK2 mutations.

A LRRK2 Gly2385Arg variant was originally reported in a PD family from Taiwan (Mata et al. 2005). Evidence for co-segregation with PD in that family was limited due to the small pedigree size though the variant was not detected in ethnically matched controls. However, in a large case-control sample of Taiwanese Chinese, this variant was recently shown to be a common polymorphism, which was significantly more frequent among PD patients than controls and therefore proposed as a risk factor for PD (Di Fonzo et al. 2006a). Interestingly, this variant has yet to be detected in Caucasians (Di Fonzo et al. 2006b).

The biological plausibility of Gly2385Arg as a pathogenic variant and its apparent ethnic specificity prompted us to conduct a case control study in an independent Chinese population and conduct the first in-vitro study examining its potential toxicity.

Methods

Ethnic Chinese subjects diagnosed with idiopathic PD by movement disorders neurologists at two major movement disorders centers (Singapore General Hospital and National Neuroscience Institute) in Singapore were included. The PD diagnosis was made in accordance with the UK PD Society Brain Bank Clinical Diagnostic Criteria of PD (Hughes et al. 1992). Sporadic PD was defined as PD without a family history of disease. The PD subjects were previously screened negative for at least 14 reported LRRK2 mutations (Tan et al. 2006). Controls of similar age, gender and race were recruited. These controls were healthy individuals without neurodegenerative diseases and examined by the authors. The institutional ethics committees approved the study and informed consent was obtained from all study subjects.

Genetic analysis

DNA was extracted from blood samples of study subjects. Mutations were screened with MALDI-TOF mass spectrometry using the Sequenom MassARRAYTM system (San Diego, CA, USA) as previously described (Tan et al. 2006). Briefly, multiplex genotyping assays were designed using the Sequenom DESIGNER software (San Diego, CA, USA). Initial PCR (5 ng of genomic DNA) and primer extension reactions were carried out according to the Sequenom MassEXTEND protocol. After purification, 15 nl of primer extension product was analyzed with a MassARRAY Sequenom-Bruker Spectrometer (Bruker Biosciences). The variant detected on the mass spectrometry was confirmed on sequence analysis. We undertook sequencing in both the forward and reverse directions according to the manufacturers’ instructions (BigDye, Applied Biosystems, Warrington, UK). Electrophoreses of the products were carried out using the ABI 3100 automated DNA sequencer (Applied Biosystems).

Construction of expression plasmids

The 9 Kb LRRK2 gene was isolated from a primary cDNA library containing the complete ORF of 7.5 Kb plus additional 5′ and 3′ UTRs in CMV vectors (TrueClone collection OriGene). The Gly2385Arg LRRK2 variant was generated by PCR-mediated site-directed mutagenesis (Stratagene) in the full-length LRRK2 cDNA. The integrity of the constructs was confirmed by DNA sequencing.

The wild type and variant LRRK2 gene were cloned into pEGFP-N1 gene in frame with EGFP coding sequences, with no intervening in-frame stop codons. The LRRK2 gene (including the initiating ATG codon) was fused to the N terminus of EGFP, and retained the fluorescent properties of the native protein.

Cell culture and transfection

Human embryonic kidney cells (HEK-293T) were maintained at 37°C in a 95% air/5% CO2 humidified atmosphere incubator and grown on polyornithine-precoated 6 well plate in Dulbecco’s modified Eagle’s medium (DMEM, GIBCO) containing 10% heat inactivated fetal bovine serum (GIBCO). Cells were transfected overnight with 3.0 μg of each plasmid DNA using LipofectamineTM 2000 reagent (Invitrogen) and OptiMEM culture medium. Cells were further incubated overnight in fresh medium containing oxygen stress reagent H2O2 (0.4 and 0.8 mM). The cultured cells were harvested by trypsin digestion for apoptosis analysis.

Apoptosis detection

For quantitation of apoptosis on H2O2 treated wild type and variant LRRK2 gene transfected cell population, Annexin V-PE apoptosis Kit I (BD Biosciences) was used according to manufacturer’s instruction. The cells were washed in cold PBS buffer twice. After the cells were resuspended in binding buffer, 100 μl of cells were transferred to a 5 ml tube and 5 μl of Annexin V-PE and 5 μl of 7-AAD were added. After incubating the reaction for 15 min in the dark, 400 μl of binding buffer were added for flow cytometry analysis (FACSCalibur, Becton Dickinson). EGFP positive cells were gated and analyzed with Annexin V and 7-AAD staining for apoptotic cells (Annexin V positive, 7-AAD negative) and late apoptotic or necrotic cells (both Annexin V and 7-ADD positive). The evaluation was carried out by an investigator blinded to the concentration of H202 and type of transfected cells. The mean value for three sets of experiments was taken for the analysis.

Determination of localization for LRRK2 variant

After overnight transfection, cells on coverslips were further incubated in fresh DMEM culture medium for 6 h and then rinsed with PBS and fixed with 3.7% formaldehyde for 10 min at RT. After washing in PBS, coverslips were mounted onto glass slides with DAPI containing Vectormount (Vector Laboratories) and examined with an LSM-510 laser-scanning confocal microscope (Carl Zeiss).

In-silico modeling

LRRK2 in silico molecular modeling was performed using the PyMolTM Molecular Graphics System 0.97 (http://www.pymol.org).

Statistical analysis

Categorical and numerical variables were analyzed with the chi-square and student t test. Genotype frequencies in the study subjects were assessed for deviation from Hardy–Weinberg equilibrium (HWE) using chi-square test for deviation from HWE. For logistic regression, the disease group (PD vs. controls) was the dependent variable, and the genotype the independent factor with adjustments made for the effect of age and gender. Statistical significance was defined at P < 0.05.

Results

A total of 989 subjects comprising of 494 PD and 495 controls similar in age and gender were analyzed. The mean age of PD and controls was 64 ± 11 and 60 ± 10 years, comprising of 56 and 54% men, respectively. Only 8.3% of cases reported a positive family history of PD or suspected parkinsonism.

The genotype frequency of the Gly2385Arg in both PD and controls followed the HWE (Table 1, P = 0.680 and P = 0.742 for controls and PD, respectively). The frequency of the heterozygous Gly2385Arg genotype was higher in PD compared to controls (7.3 vs. 3.6%, odds ratio = 2.1, 95% CI: 1.1–3.9, P = 0.014).
Table 1

Gly2385Arg genotype frequency in PD and controls

 

Controls

PD

Total

Genotype

 GG

476 (96.36%)

458 (92.53%)

934 (94.44%)

 GA

18 (3.64%)

36 (7.27%)

54 (5.46%)

 AA

0

1 (0.2%)

1 (0.1%)

Total

494

495

989

Using this observed value of OR as an estimate of relative risk (RR), and the frequency of the risk genotype (heterozygous Gly2385Arg) among controls as an estimate of its frequency in the general population, yields a population attributable risk (PAR) of ∼4% for the Gly2385Arg heterozygous genotype.

In the multivariate logistic regression analysis with the disease group (PD vs. controls) as the dependent variable and genotype as an independent factor with adjustments made for the effect of age and gender, we found that the heterozygous Gly2385Arg genotype was associated with an increased risk of PD compared to controls (odds ratio = 2.67, 95% CI: 1.43–4.99, P = 0.002). This association remained significant even if we included only those PD patients without a family history (odds ratio = 2.47, 95% CI: 1.29–4.72, P = 0.006).

Confocal microscopy revealed the localization of the Gly2385Arg variant was primarily cytoplasmic with some aggregates, similar to wild type protein (Fig. 1). There was a higher rate of apoptosis when the Gly2385Arg variant was over-expressed in human HEK cell lines compared to wild type protein under a higher concentration of H202 (0.8 mM) (Fig. 2). The transfection efficiency in the cell lines was similar for both wild type and variant. For 0.8 mM H202 exposure, the mean percentage of apoptosis, necrosis and cell death observed for the Gly2385Arg variant transfected cells compared to wild type was 8.5 ± 1.7, versus 2.9 ± 1.3, P = 0.01, 8.1 ± 2.3, versus 6.9 ± 1.0, P = 0.43, 16.6 ± 0.66, versus 9.8 ± 2.3, P = 0.007, respectively. For the 0.4 mM H202 exposure, the mean percentage of apoptosis, necrosis and cell death observed for the Gly2385Arg variant transfected cells compared to wild type was 1.2 ± 0.3, versus 0.9 ± 0.3, P = 0.37, 3.3 ± 0.2, versus 3.3 ± 0.7, P = 0.9, 4.6 ± 0.2, versus 4.3 ± 1.1, P = 0.7, respectively. No difference was observed at baseline without H2O2 exposure. We were unable to consistently reproduce the findings using human SH-HY5Y cell lines due to the low transfection efficiency we experienced using this cell line. We found the expression of LRRK2 to be 3.18 times higher in HEK than SH-SY5Y cells. The ratio of expression of Gly2385Arg variant:wild-type in HEK and SH-SY5Y was 1.28 and 0.22. The expression of the Gly2385Arg variant was three times higher in HEK than in SH-SY5Y cell lines.
https://static-content.springer.com/image/art%3A10.1007%2Fs00439-006-0268-0/MediaObjects/439_2006_268_Fig1_HTML.jpg
Fig. 1

Superimposed confocal image of GFP (green) and DAPI (blue). Bars denote 10 μm. Both wildtype and Gly2385Arg proteins exhibit similar cytoplasmic distribution and protein aggregations

https://static-content.springer.com/image/art%3A10.1007%2Fs00439-006-0268-0/MediaObjects/439_2006_268_Fig2_HTML.gif
Fig. 2

Rate of apotoptic, necrotic and overall cell death after exposure to different concentration of hydrogen peroxide

Discussion

Mutations in the LRRK2 gene are the most frequent known cause of familial PD. However, due to incomplete penetrance, they are also found in few, apparently sporadic PD cases. Whether LRRK2 polymorphic variants influence the susceptibility to the common forms of PD remains controversial. There is still little evidence thus far that common LRRK2 variants are associated with PD in Caucasians (Biskup et al. 2005; Paisan-Ruiz et al. 2006; Paisan-Ruiz et al. 2005). However, we obtained evidence that LRRK2 haplotypes were associated with PD in the Chinese population from Singapore (Skipper et al. 2005b), and a common LRRK2 missense variant, Gly2385Arg, was independently detected as a putative risk factor for PD in the Chinese population from Taiwan (Di Fonzo et al. 2006a). Allelic association studies are prone to false positive results, mainly due to small sample sizes, population stratification and genotyping errors. Replication of the association in independent, large samples is therefore crucial in order to validate the results.

In the first part of our study, we demonstrated that the Gly2385Arg heterozygote variant was associated with an increased risk of PD. In the multivariate analysis, subjects with the variant were 2.5 times more likely to develop PD. Our findings in an independent cohort of Chinese population of different geographical location confirmed the initial observation that the variant is a common polymorphism in Chinese and it was a risk factor in this race (Di Fonzo et al. 2006a). Furthermore, a meta-analysis of our current study and previous study in Taiwanese (Di Fonzo et al. 2006a) revealed no heterogeneity of the genotypes in both cohorts and the overall odds ratio was 2.2 (95% CI 1.5–3.3) and the association involving about 2000 study subjects was significant (PD vs. controls, 9.6 vs. 4.3%, P = 0.001). These odds ratios are in the range that one can expect for the risk factors of a disease with a complex causation, in which multiple gene–gene and gene–environmental interactions might also play important roles.

We used the observed frequency of the risk genotype (heterozygous Gly2385Arg) among controls and the observed value of OR as estimates of the risk genotype frequency in the general population, and of the RR, respectively. Under these assumptions, a PAR of ∼4% for the Gly2385Arg heterozygous genotype is obtained, delineating this variant as the most important genetic risk factor for PD identified so far in the Chinese population, where the Gly2019Ser LRRK2 mutation appears absent.

It is also interesting that the Gly2385Arg variant was not detected in any previous screen for LRRK2 mutations in Caucasians, including one study, in which a total of 220 chromosomes (PD and controls) were sequenced (Di Fonzo et al. 2006b). This variant appears therefore specific for the Asian population, emphasizing further the fact that ethnic-specific effects should be considered in genetic association studies.

The potential functional significance of this variant which lies in the LRRK2 WD domain is supported by a number of observations. The WD40 domain is found in a number of eukaryotic proteins which are responsible for a variety of functions such as signal transduction, pre-mRNA processing and cytoskeleton assembly. The G residue is hydrophobic while R has positive charge and hydrophilic. According to the PROSITE (Hulo et al. 2006) the motifs from 2385 to 2390 (GLidCV) is one of the 25 potential N-myristoylation site of LRRK2 protein. The position 2385 G can form a N-myristoylation site in the LRRK2 protein. Posttranslational N-myristoylation is required for the anti-apoptotic activity of some human proteins, such as gelsolin (Sakurai and Utsumi 2006) and myristoylation of caspase-activated p21-activated protein kinase 2 potentiates late apoptotic events (Vilas et al. 2006). Myristoylation can also be a critical step triggering the apoptosis event. The BID protein was reported to induce the release of cytochrome C from mitochondria and subsequently initiate cell death when it was myristoylated (Zha et al. 2000).

The Gly2385 residue is not conserved in the evolution of LRRK2 proteins, but residues with a net positive charge, such as Arginine, are not observed in the known LRRK2 homologs (Di Fonzo et al. 2006a). Intriguingly, the WD40 domain of LRRK2 has a high net positive charge, due to the presence of several Arginine residues. Structural simulation studies predict that the Gly2385 residue is exposed at the surface of the WD40 domain, where it increases the net positive charge of the domain, with possible important functional implications for the biology of this domain (Mata et al. 2006). The arginine could either prevent protein interactors that might bind with glycine or leads to new interactions with other proteins as it is on the outside of the molecule where interactions could occur (see Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00439-006-0268-0/MediaObjects/439_2006_268_Fig3_HTML.jpg
Fig. 3

The predicted 3D structure of part of the LRRK2 WD40 domain. The Arginine 2385 is exposed at the surface of the domain, where it introduces an additional net positive electric charge

Under confocal microscopy, we demonstrated that the cytoplasmic localization of the variant appears similar to wild type protein, with presence of protein aggregations. Similar localization pattern has also been observed with other pathogenic mutations (such as the G2019S). Previous studies on PD risk alleles (such as Parkin promoter variant) in cell models have demonstrated that the cellular toxic effect of such alleles could be “amplified” in the presence of oxidative stress environment (Tan et al. 2005b). Our in-vitro study using hydrogen peroxide in a cellular stress model revealed that the variant was associated with a greater rate of apoptosis and cell death only under higher oxidative exposure (0.8 mM), while there was little effect with a lower exposure (0.4 mM). These observations corroborate clinical observation that this common variant (which is also present in healthy individuals) is probably a risk allele and may become pathogenic when internal or external environmental stress is present. As our finding was consistently reproduced in only one human cell line in-vitro, in-vivo studies are needed to validate and further address the pathogenicity of this variant.

Oxidative stress and ubiquitin proteasomal impairment are well-recognized features in various cellular and animal PD models and post-mortem human PD specimens (Abou-Sleiman et al. 2006). The generation of reactive oxygen species within neuronal cells may be modulated by both internal and external environmental stimuli. This can result in triggering specific signaling cascades, such as mitogen-activated protein kinases (MAPK), which can determine cell survival or death. The regulatory mechanisms of LRRK2 and its involvement in oxidative stress-induced apoptosis have not been clarified. In cellular models, LRRK2 mutations are associated with an increased kinase activity (West et al. 2005; Gloeckner et al. 2006). It is possible that mutations or risk variants outside the MAPKKK domain may exert their influence by regulating the kinase domain. However, as the physiologic substrates of LRRK2 are still not known, further studies are needed to investigate whether the presence of the arginine residue in the Gly2385Arg modulates the interaction with these substrates, the capability of LRRK2 to form homo-dimers or directly influences the kinase activity of the protein.

In conclusion, our study lends support to the contention that the Gly2385Arg variant is a common risk factor for PD in the Chinese population, with an estimated PAR of ∼4%. In addition, we have also shown in bioinformatics and in-vitro studies that the Gly2385Arg variant is biologically relevant and it might act through pro-apoptotic mechanisms. Our findings have important implications for understanding the mechanisms of PD. Further experiments to unravel the pathophysiologic significance of Gly2385Arg may also be valuable for defining at-risk individuals or more homogeneous subgroups of patients for clinical trials.

Acknowledgments

The study was supported by the National Medical Research Council, Genome Institute of Singapore, Biomedical Research Council, Department of Clinical Research, Singapore General Hospital. We thank Drs Liu Ed, Lee WL and Aw SE for their support.

Copyright information

© Springer-Verlag 2006