Kinase activity is required for the toxic effects of mutant LRRK2/dardarin

Abstract

Mutations in the LRRK2 gene, coding for dardarin, cause dominantly inherited Parkinson's disease (PD). Dardarin is a large protein, and mutations are found throughout the gene including the kinase domain. However, it is not clear if kinase activity is important for the damaging effects of pathogenic mutations. In this study, we noted two cellular phenotypes associated with mutant dardarin. First, pathogenic mutations increase the tendency of dardarin to form inclusion bodies. Secondly, neurons and neuronal cell lines undergo cell death after expression of mutant protein. Manipulating activity by replacing the kinase domain with a ‘kinase-dead’ version blocks inclusion body formation and strongly delays cell death. This predicts that kinase inhibitors will be useful therapeutic agents in patients with LRRK2 mutations and, perhaps, in sporadic PD. We also show that dardarin protein is expressed within human midbrain neurons and that C-terminal epitopes are also found in some Lewy bodies.

Introduction

Dominant mutations in the LRRK2 gene, coding for the protein dardarin (Paisan-Ruiz et al., 2004, Zimprich et al., 2004), are a common cause of inherited Parkinson disease (PD), accounting for 1–6% of cases including many apparently sporadic cases (Cookson et al., 2005). Patients with LRRK2/dardarin mutations show a loss of striatal ¹⁸F-DOPA PET signal, consistent with a loss of presynaptic neurons projecting from the substantia nigra pars compacta (SNpc) to the striatum (Hernandez et al., 2005). These findings correlate well with post-mortem examination of PARK8 cases (Funayama et al., 2002, Zimprich et al., 2004), which revealed loss of nigral neurons, usually with the presence of α-synuclein positive Lewy bodies (Zimprich et al., 2004). However, the phenotype of cases with LRRK2 mutations also includes deposits of the microtubule-associated protein tau and several cases have intracellular protein inclusions that are tau and α-synuclein-negative (Zimprich et al., 2004). Some cases have apparently ‘pure’ nigral degeneration in the absence of Lewy bodies, and others may have symptoms such as dementia and amyotrophy that indicate damage to regions of the brain outside of the nigra (Funayama et al., 2002, Zimprich et al., 2004). These observations demonstrate that dominant mutations are associated with cell loss in the SNpc, presumably the anatomical substrate of the parkinsonism in these patients, and other brain regions together with a variable protein inclusion.

The dardarin protein is a large, 2527 amino acid protein with multiple discrete domains. The N-terminal region is followed by several leucine-rich repeats. In the central region, there is a GTPase domain (also known as Roc for Ras of complex proteins) that is separated from a kinase domain by a region referred to as the COR (C-terminal of Roc) domain (Bosgraaf and Van Haastert, 2003). Finally, the carboxy terminal region has a WD40-like domain that may form a β-propeller structure (Fig. 1A). This complex domain structure is notable for two reasons. Firstly, dardarin is likely to have kinase and GTPase activities, which may be linked. One hypothesis is that the GTPase activity might control kinase activity (Shen, 2004), which has recently been demonstrated to be true for the paralagous kinase LRRK1 (Korr et al., 2006). A common mutation in LRRK2/dardarin (G2019S) is found in the Mg²⁺ binding motif at the beginning of the activation loop of the kinase domain, while a second mutation at the adjacent isoleucine (I2020T) is found in the original family reported to have autosomal dominant PD linked to this locus (Funayama et al., 2005). Secondly, regions such as the leucine-rich repeats (LRRs), Roc, COR or WD40 domains may be capable of mediating protein–protein interactions. Mutations are found in all of these regions, including I1122V in the LRRs, R1441C in the Roc domain and Y1699C in the COR region.

It is not clear how mutations in dardarin lead to dominantly inherited disease that include both cell loss and variable protein inclusion pathology. For example, it has been proposed that kinase domain mutations might increase (Toft et al., 2005) or decrease (Albrecht, 2005) kinase activity. By analogy with mutations in presenilin, which increase processing of the amyloid precursor protein to the toxic Aβ peptides (Brunkan and Goate, 2005), a hyperactive kinase might increase phosphorylation of its target(s) leading to increased cellular damage. Indeed, recent data support the hypothesis that dardarin mutations in the kinase domain, G2019S and I2020T, increase kinase activity (Gloeckner et al., 2006, West et al., 2005), although the analogous mutations in LRRK1 actually decrease activity (Korr et al., 2006). The one mutation in dardarin (R1441C) outside of the kinase domain that has been assayed exerts only modest effects on activity (West et al., 2005). Whether there is a consistent effect of mutations on kinase activity is therefore currently unclear.

However, observations of increased activity would not be proof that kinase activity is required for pathogenesis as other properties of the mutant protein than kinase activity may also be altered. An analogy would be mutations in Cu/Zn superoxide dismutase (SOD1), associated with familial amyotrophic lateral sclerosis. These mutations have variable effects on dismutase activity but consistent effects on the tendency of the mutant protein to aggregate (Valentine and Hart, 2003), and therefore although some mutations do alter enzyme activity, pathogenesis correlates with another, novel, gain of function.

One way to distinguish between these possibilities is to measure a dominant phenotype and establish if this is prevented when kinase activity is decreased. To do this, one would need an indirect assay of the dominant mutations other than kinase activity itself. In this study, we show that mutations in dardarin lead to two phenotypes in cells, inclusion body formation and cell death, as recently suggested elsewhere (Smith et al., 2005). We support the idea that these two phenotypes are relevant to the human disease by showing that the protein is expressed in dopaminergic neurons in midbrain melanized neurons from human post-mortem material. Furthermore, we show that the cellular phenotypes are ameliorated by genetically inactivating the kinase using mutagenesis of key residues, suggesting that kinase activity is required for the full dominant effects of mutant dardarin.