Human Molecular Genetics Advance Access originally published online on May 19, 2006
Human Molecular Genetics 2006 15(13):2059-2075; doi:10.1093/hmg/ddl131
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Biochemical analysis of Parkinson's disease-causing variants of Parkin, an E3 ubiquitin–protein ligase with monoubiquitylation capacity
1 Neurologie et Thérapeutique Expérimentale, INSERM U679-Université Pierre & Marie Curie, Paris VI, Hôpital de la Pitié-Salpêtrière, 75013 Paris, France, 2 Département de Génétique, Cytogénétique et Embryologie, Hôpital de la Pitié-Salpêtrière, AP-HP, Paris, France and 3 Fédération de Neurologie, Hôpital de la Pitié-Salpêtrière, AP-HP, Paris, France
* To whom correspondence should be addressed. Tel: +33 142162217; fax: +33 144243658; Email: corti{at}ccr.jussieu.fr
Received January 26, 2006; Revised April 4, 2006; Accepted May 12, 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Mutations in the parkin gene, encoding an E3 ubiquitin–protein ligase, are a frequent cause of autosomal recessive parkinsonism and are also involved in sporadic Parkinson's disease. Loss of Parkin function is thought to compromise the polyubiquitylation and proteasomal degradation of specific substrates, leading to their deleterious accumulation. Several studies have analyzed the effects of parkin gene mutations on the biochemical properties of the protein. However, the absence of a cell-free system for studying intrinsic Parkin activity has limited the interpretation of these studies. Here we describe the biochemical characterization of Parkin and 10 pathogenic variants carrying amino-acid substitutions throughout the sequence. Mutations in the RING fingers or the ubiquitin-like domain decreased the solubility of the protein in detergent and increased its tendency to form visible aggregates. None of the mutations studied compromised the binding of Parkin to a series of known protein partners/substrates. Moreover, only two variants with substitutions of conserved cysteine residues of the second RING finger were inactive in a purely in vitro ubiquitylation assay, demonstrating that loss of ligase activity is a minor pathogenic mechanism. Interestingly, in this in vitro assay, Parkin catalyzed the linkage of single ubiquitin molecules only, whereas the ubiquitin–protein ligases CHIP and Mdm2 promoted the formation of polyubiquitin chains. Similarly, in mammalian cells Parkin promoted the multimonoubiquitylation of its substrate p38, rather than its polyubiquitylation. Thus, Parkin may mediate polyubiquitylation or proteasome-independent monoubiquitylation depending on the protein context. The discovery of monoubiquitylated Parkin species in cells hints at a novel post-translational modification potentially involved in the regulation of Parkin function.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Parkinson's disease (PD) is the second most common neurodegenerative disorder after Alzheimer's disease, affecting nearly 2% of individuals over the age of 65 (1). Clinically, it is characterized by a triad of invalidating neurological symptoms—resting tremor, rigidity and bradykinesia—due to the progressive, preferential degeneration of the dopaminergic neurons of the substantia nigra pars compacta. Lewy bodies (LBs), ubiquitylated intraneuronal inclusions enriched in
-synuclein, are a neuropathological hallmark invariably associated with idiopathic PD (2).
Although the etiology of sporadic PD is poorly understood, there is evidence that both environmental factors and genetic predisposition contribute to its development. Rare familial forms of the disease exist: they are compatible with an autosomal dominant or recessive mode of inheritance. Since 1997, five genes unambiguously linked to these diseases have been identified. Rare missense mutations and more frequent multiplications of a large genomic region including the -synuclein gene cause autosomal dominant parkinsonian syndromes (3). These syndromes are most frequently due to mutations in the recently identified LRRK2 gene, with one single mutation—G2019S—accounting for 3–40% of all cases of autosomal dominant parkinsonism, depending on the population studied (4–11).
Mutations in three additional genes, parkin, DJ-1 and PINK1, cause autosomal recessive parkinsonism, usually with early onset (3). A series of parkin exon rearrangements (deletions/multiplications) and point mutations (missense/truncating) affecting the entire coding sequence are responsible for nearly half of all cases of autosomal recessive parkinsonism with early onset, in populations of various ethnic origins (12–14). Parkin gene mutations also account for more than 15% of sporadic PD cases with early onset (15). These mutations are associated with ages at onset varying from 7 to 72 years and a broad phenotypic spectrum, including cases similar to dopa-responsive dystonia, with atypical signs or resembling idiopathic PD (13,16–18). In terms of neuropathology, LBs are absent in parkin-linked parkinsonism, but LB pathology and/or tau deposits or -synuclein immunoreactive inclusions have been reported (17,19–24). Despite this phenotypic variability, there is no clear relationship between the nature and position of the mutation and the clinical severity of the disease (25). Nevertheless, it has been suggested that amino-acid substitutions within functional domains lead to earlier disease onset when compared with substitutions in domains with unknown function (25).
Since the discovery of the E3 ubiquitin–protein ligase activity of Parkin in 2000, it has been suggested that Parkin targets specific protein substrates for proteasomal degradation (26). It is currently thought that the loss of this function, due to disease-causing parkin gene mutations, leads to the abnormal accumulation of toxic proteins and neurodegeneration. The mechanisms by which parkin gene mutations lead to a loss of protein function remain a matter of debate. The existence of a series of severe truncating mutations and the lack of correlation between the type of mutation and the clinical presentation of the disease are consistent with the complete loss of E3 ubiquitin–protein ligase activity being the major pathogenic mechanism. However, it is difficult to conceive that a series of substitutions located within or outside essential functional Parkin domains systematically lead to loss of catalytic activity. Studies from several laboratories, including ours, have suggested that the aggregation of Parkin or its pathogenic variants, either spontaneously or under stress conditions, leads to mislocalization and/or dysfunction of the protein in parkin-related and potentially sporadic PD (27–37). However, the lack of a purely in vitro assay for assessing Parkin enzymatic activity has complicated the biochemical analysis of parkin gene mutations. In a semi-in vitro ubiquitylation assay, Shimura et al. (26) demonstrated a complete loss of the ubiquitylation activity of Parkin variants carrying single pathogenic amino-acid changes. Various groups have since explored the ability of selected pathogenic Parkin variants to promote their own ubiquitylation, and that of cellular proteins or of particular substrates (32,38–48). These studies were performed in vivo, in transfected cells, or in vitro, using immunoprecipitated or in vitro-translated Parkin. In these cell-derived systems, contaminating E3 ligases or cofactors of ubiquitylation reactions may interfere with or modify Parkin activity, making results difficult to interpret. These studies have generated conflicting results, with some showing a loss of ubiquitin–protein ligase activity and others its preservation, for one and the same mutation (26,32,40,41,43,46,48).
Another unresolved issue in this field concerns the type of ubiquitin modification induced by Parkin. Theoretically, proteins may be ubiquitylated by the attachment of (i) a single ubiquitin molecule to one lysine residue (monoubiquitylation); (ii) single ubiquitin molecules to multiple lysine residues (multimonoubiquitylation); or (iii) multimeric chains of ubiquitin (polyubiquitylation) (49,50). Polyubiquitin chains may also be assembled through isopeptide bonds involving specific lysine residues in ubiquitin. Ubiquitin chains linked via Lys-48 are targeted for proteasomal degradation. In contrast, the use of Lys-63 for polyubiquitin chain assembly and the conjugation of single ubiquitin molecules to substrates correspond to post-translational modifications involved in diverse cellular processes, including the endocytosis of membrane proteins, protein sorting and trafficking, and transcriptional regulation (49,50). At least 10 putative Parkin substrates have been identified. These substrates are involved in processes as diverse as cell signaling (Pael-R), cell-cycle control (cyclin E), protein biosynthesis (p38 subunit of aminoacyl-tRNA synthetase complexes), cytoskeletal dynamics (-/ß-tubulin), neurotransmitter uptake (dopamine transporter), synaptic vesicle release (CDCrel-1/2, synaptotagmin XI) and other synaptic functions (O-glycosylated -synuclein, synphilin) (38–46,51). Parkin has been shown to ubiquitylate and to accelerate the degradation of several of these proteins in transfected cells. In addition, the levels of a few of these substrates have been shown to be moderately higher in the brains of patients with parkin gene mutations (39,40,45,48,51) and, in some cases, immunoreactivity to these substrates has been reported in LBs, in cases of idiopathic PD (41,42,52–54). It remains unclear whether any of these proteins cause dopaminergic neuronal death in parkin-related parkinsonism, but these findings are consistent with Parkin playing a major role in the polyubiquitylation and proteasomal degradation of cellular proteins. Accordingly, Parkin has been shown to promote Lys-48-linked polyubiquitylation in cell-derived systems (55,56). However, Parkin may also mediate proteasomal-independent, Lys-63-linked polyubiquitylation (55,56), and recent data indicate a possible role for this protein in monoubiquitylation (57).
We investigated the intrinsic biochemical properties of Parkin and the detrimental consequences of parkin gene mutations further, by detailed characterization of Parkin and a series of 10 pathogenic variants carrying amino-acid substitutions throughout the protein sequence in cells and in vitro, using a reconstituted cell-free system to study the catalytic activity of the protein.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Missense parkin gene mutations frequently lead to changes in the detergent solubility of Parkin
The Triton X-100 solubility of a series of 10 pathogenic Parkin variants carrying amino-acid substitutions in the various functional domains was compared with that of normal Parkin, after overproduction of the corresponding N-terminally HA-tagged proteins in COS7 cells (Fig. 1). A summary of the genetic data supporting the pathogenic role of each of the chosen substitutions is provided in Table 1. Qualitative western blot analysis of the distribution of Parkin in the detergent-soluble and pellet fractions with anti-HA antibodies revealed that the solubility of some of the variants studied (A82E, K161N, K211N, R256C, G328E) was similar to that of normal Parkin, whereas other variants (R42P, R275W, C289G, C418R, C441R), mostly those with amino-acid substitutions in the C-terminal RING-fingers of the protein, accumulated in large amounts in the detergent-insoluble pellets (Fig. 1A and B). Similar patterns of solubility were observed when anti-Parkin antibodies were used for western blotting (Fig. 1C). In contrast to anti-HA antibodies, which only recognize a full-length protein of an apparent molecular mass of 52 kDa, anti-Parkin antibodies also detect an N-terminally truncated isoform previously described in the literature (34,58). In general, the solubility of the truncated isoform was similar to that of the full-length protein: it was rather soluble in the case of normal Parkin and its A82E, K161N, K211N, R256C and G328E variants, whereas it tended to be at least as insoluble as the full-length protein in the case of the C289G, C418R and C441R variants. However, in the case of the R42P and R275W variants, the truncated isoform was consistently more soluble than the 52 kDa protein. This configuration was expected for the R42P variant, as its truncated isoform does not carry the corresponding amino-acid substitution, but was surprising for the RING1 R275W variant. Semi-quantitative analysis of the relative levels of the full-length protein by densitometry confirmed the qualitative results (Fig. 1D). It also revealed that the R256C variant was significantly less soluble in Triton X-100 than normal Parkin, whereas the A82E variant tended to be more soluble. In addition, the relative protein levels of the R42P, C418R and C441R variants in total cell extracts tended to be lower than those of the other variants, suggesting that these variants had shorter half-lives (Fig. 1E). We analyzed the intracellular distribution of Parkin variants by immunocytochemistry using anti-Parkin antibodies, as means of correlating the detergent solubility and the ability to form visible aggregates in COS7 cells. Cells containing aggregates were frequently observed following overproduction of the R275W, C289G, C418R and C441R variants (Fig. 2 and data not shown). In contrast, the R42P and R256C variants behaved similar to normal Parkin and the A82E, K161N, K211N, R256C and G328E variants, which only rarely formed aggregates in transfected cells (Fig. 2 and data not shown).
|
|
|
Missense parkin gene mutations do not compromise the ability of Parkin to interact with known protein partners or substrates
We addressed the ability of Parkin variants carrying pathogenic amino-acid substitutions to interact physically with known protein partners or substrates by GST pull-down. To this end, we optimized the conditions for the production of a soluble GST–Parkin fusion protein able to reproduce previously reported protein interactions. When produced in Escherichia coli at 30 or 37°C, most of the protein were sequestered in inclusion bodies and nearly 100 µg of soluble protein was recovered on glutathione-sepharose beads per liter of bacterial culture. However, reproducible protein interactions and enzymatic activity (see paragraph below) were only observed when GST–Parkin was produced at 30°C and IPTG induction times were kept below 2 h. As illustrated in Figure 3A with the examples of Hsp70 and
-tubulin, when bacteria were grown at 37°C and IPTG induction times were longer than 2 h, the interaction of soluble GST–Parkin with its protein partners was weak and unreliable. On the basis of these observations, GST–Parkin and GST–pathogenic variant fusion proteins were produced in the appropriate conditions and affinity-purified on glutathione-sepharose beads. Similar amounts of soluble protein were recovered in each case (Fig. 3B). Equal amounts of recombinant proteins were incubated with cell lysates from native COS7 cells or cells overproducing -synuclein, or Parkin substrates CDCrel-1 or p38. The binding of Parkin variants to exogenous -synuclein, CDCrel-1 and p38, and to endogenous - and 
-tubulin, Hsp70 and the proteasomal 4 subunit was analyzed by western blotting using specific antibodies (Fig. 3C). As expected, GST–Parkin interacted with CDCrel-1, p38, - and -tubulin, Hsp70 and the proteasomal 4 subunit, but not with -synuclein. These interactions were specific, as GST did not interact with any of the indicated proteins. None of the pathogenic amino-acid substitutions tested compromised the ability of Parkin to interact with the proteins studied.
|
Abolition of the E3 ubiquitin–protein ligase activity of Parkin is a rare consequence of missense parkin gene mutations
We developed a purely in vitro Parkin-dependent autoubiquitylation assay for assessing the intrinsic E3 ubiquitin–protein ligase activity of Parkin. GST–Parkin affinity-purified from E. coli extracts on glutathione-sepharose beads was eluted from the beads and incubated with recombinant, bacterially produced E1 ubiquitin-activating enzyme, E2 ubiquitin-conjugating enzyme UbcH7 and FLAG-tagged ubiquitin, in the presence of ATP (Fig. 4). GST–Parkin was then repurified on glutathione-sepharose beads and examined by western blotting using anti-FLAG antibodies after migration on a 15% denaturing polyacrylamide gel (Fig. 4A, left panel). The supernatants containing the remaining components of the ubiquitylation machinery were also migrated on a 15% gel which was stained with Coomassie blue for detection of ubiquitin and E2 (Fig. 4A, right panel). For better resolution of the high molecular weight ubiquitylated protein species, aliquots of beads and supernatants were migrated on a 3–8% gradient gel and analyzed by anti-FLAG immunoblotting (Fig. 4B). Two major ubiquitylated protein species and, after longer exposure times, several minor ubiquitin-positive protein species of higher apparent molecular mass were observed exclusively in the bead fraction, demonstrating the efficient autoubiquitylation of GST–Parkin. In the absence of ATP, E1, E2, ubiquitin or GST–Parkin, no ubiquitylated protein species were detected. Similar results were obtained if a mixture of Ubc6 and Ubc7 enzymes was used in place of UbcH7 (data not shown). The total amount of FLAG-positive protein produced in the test was proportional to the amount of E3 ubiquitin–protein ligase added (Fig. 4C) and the reaction time (Fig. 4D).
|
We cleaved the N-terminal GST moiety, using PreScission protease, to determine whether it interferes with the E3 ubiquitin–protein ligase activity of Parkin. We then analyzed and compared the autoubiquitylation of GST–Parkin and Parkin (Fig. 4E). The pattern of ubiquitylation and the intensity of signal were similar for the two proteins. We therefore used GST fusion proteins to evaluate the effect of missense parkin gene mutations on the E3 ubiquitin–protein ligase activity of Parkin (Fig. 5A). In vitro analysis of the E3 ubiquitin–protein ligase activity of these Parkin variants revealed that most promoted their own ubiquitylation as efficiently as normal Parkin. Only two of the pathogenic variants analyzed, Parkin-C418R and Parkin-C441R, lacked E3 ubiquitin–protein ligase activity. These variants carry substitutions of two critical, highly conserved cysteine residues in the second C-terminal RING domain of Parkin (Fig. 1A, Table 1). In addition, the artificial Parkin77–465 variant, lacking the N-terminal ubiquitin-like domain, displayed only weak autoubiquitylation in this assay (Fig. 5A). To confirm these results using an alternative approach, we overproduced N-terminally HA-tagged Parkin or Parkin variants in COS7 cells. After immunoprecipitation with anti-HA antibodies, the immune complexes were washed either stringently, to dissociate proteins binding to Parkin, or less stringently, to preserve protein interactions, and incubated with recombinant E1, UbcH7 and FLAG-tagged ubiquitin, as previously described for GST–Parkin. In the presence of ATP, ubiquitylated FLAG-immunoreactive protein species were observed with complexes containing normal Parkin or the R42P, A82E, K161N, K211N, R256C, R275W, C289G and G328E variants, irrespective of the washing conditions used. When stringent conditions were used to wash the immune complexes (Fig. 5B), ubiquitylated protein species were not observed with the C418R and C441R variants, confirming our previous observations with the corresponding GST–Parkin proteins. In contrast, when the washing procedure was less stringent, ubiquitylated proteins formed even in the presence of these inactive variants (data not shown), suggesting that proteins with E3 ubiquitin–protein ligase activity binding to Parkin contaminate the reaction in these conditions and account for a significant proportion of the activity measured.
|
We investigated the possibility of inactive Parkin variants affecting the E3 ubiquitin–protein ligase activity of normal Parkin by a dominant-negative mechanism, by incubating recombinant GST–Parkin alone or with GST-HA-Parkin-C418R or GST-HA-Parkin-C441R, and monitoring autoubiquitylation (Fig. 5C). Similar amounts of ubiquitylated proteins were produced when GST–Parkin was incubated alone or in the presence of either of the inactive variants.
Parkin promotes monoubiquitylation at multiple lysines in vitro
We performed autoubiquitylation assays using recombinant GST–Parkin, GST–Mdm2 or HA–CHIP in the presence of FLAG-ubiquitin to determine the type of modification induced by Parkin, for comparison with other known E3 ubiquitin–protein ligases. Autoubiquitylated protein species were analyzed by western blotting using monoclonal antibodies directed against mono- and polyubiquitylated proteins (FK2), monoclonal antibodies recognizing polyubiquitin chains only (FK1) or anti-FLAG antibodies (Fig. 6A). As expected, both FK2 and anti-FLAG antibodies recognized a smear of ubiquitylated Parkin, Mdm2 and CHIP protein species (left and right panels). FK1 also efficiently detected high molecular weight polyubiquitylated Mdm2 and CHIP species. In contrast, no signal corresponding to the presence of polyubiquitylated Parkin species was obtained with FK1, even after long exposure times (middle panel), indicating that Parkin was mostly modified by the attachment of single ubiquitin molecules to multiple lysine residues. To exclude the possibility that the lack of recognition of autoubiquitylated Parkin species by FK1 reflects steric hindrance because of the presence of the FLAG-tag on the ubiquitin molecule, rather than the absence of polyubiquitylated species, we repeated the experiment using Parkin and CHIP and untagged ubiquitin (Fig. 6B). As in our previous observations, whereas untagged ubiquitylated CHIP species were readily detected with both FK1 and FK2 antibodies, untagged ubiquitylated Parkin species were only recognized by FK2 antibodies.
|
We investigated this issue further by in vitro autoubiquitylation assays with GST–Parkin, GST–Mdm2 and HA–CHIP, in the presence of normal ubiquitin or a lysine-less ubiquitin derivative (Ub-K0) in which all seven lysine residues had been replaced by arginine (Fig. 6C). This derivative can be used for mono- and multimonoubiquitylation, which depend on the presence of the C-terminal glycine residue of ubiquitin, but lacks the conjugation sites required for polyubiquitylation. Autoubiquitylation patterns and signal intensities for GST–Parkin were similar in the presence of ubiquitin or Ub-K0, as revealed by anti-ubiquitin antibodies. These results suggest that Parkin has intrinsic mono-/multimonoubiquitylation capacity in vitro. In contrast, no signal corresponding to mono- or multimonoubiquitylated protein species was observed when GST–Mdm2 was incubated with Ub-K0, demonstrating that GST–Mdm2 is modified principally by polyubiquitylation, and suggesting that monoubiquitylated GST–Mdm2 species are highly unstable intermediates of the polyubiquitylation reaction. An intermediate situation was observed for HA–CHIP, for which a ladder of ubiquitin-immunoreactive bands was obtained in the presence of Ub-K0, suggesting that the generation of mono-/multimonoubiquitylated species accompanies the formation of polyubiquitylated proteins. This finding is consistent with the pattern of ubiquitylation observed for HA–CHIP with anti-FLAG and FK2 antibodies, showing both lower molecular weight bands and a high-molecular weight smear (Fig. 6A and B).
We have previously demonstrated that Parkin promotes the ubiquitylation of its substrate p38 in cells (41). To determine the type of ubiquitylation by which p38 was modified, we overexpressed a myc-tagged version of this protein together with Parkin and normal His-tagged ubiquitin or its lysine-less derivative in COS7 cells. His-tagged ubiquitylated proteins were purified in denaturing conditions using a nickel-charged affinity matrix and the ubiquitylation pattern of p38 was analyzed by western blotting using anti-myc antibodies (Fig. 7). The patterns of ubiquitylation observed in the presence of normal or lysine-less ubiquitin were substantially similar, supporting the idea that p38 is mainly modified by multimonoubiquitylation rather then polyubiquitylation in vivo.
|
Finally, we investigated the possible existence of mono-/multimonoubiquitylated Parkin species in COS7 cells, by overproducing HA-tagged ubiquitin alone or with FLAG-tagged Parkin. The cells were then lysed and Parkin was immunoprecipitated with specific polyclonal antibodies and analyzed by western blotting, using monoclonal FK2, anti-HA, or FK1 antibodies (Fig. 8A). Ubiquitylated Parkin species were efficiently detected with FK2 and anti-HA antibodies (left panels). However, as previously observed in vitro, these species were not recognized by FK-1 antibodies unless proteasome activity was inhibited (right panels), although these antibodies readily recognized polyubiquitylated proteins in cell lysates. We further explored the type of ubiquitylation by which Parkin was modified by overexpressing FLAG-tagged Parkin with His-tagged normal or lysine-less ubiquitin in COS7 cells (Fig. 8B). After purification of His-tagged ubiquitylated proteins, the ubiquitylation of Parkin was followed by western blotting using anti-FLAG antibodies. The patterns of Parkin ubiquitylation observed with normal and lysine-less ubiquitin largely overlapped, except for the presence of a more abundant higher molecular-weight smear with normal ubiquitin. Thus, in basal conditions, a fraction of Parkin is modified by multimonoubiquitylation in cells.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Various attempts have been made to explore the functional consequences of parkin gene mutations, but no unifying clear-cut mechanism of loss of Parkin function has yet been identified (26,30–35,38). Moreover, the intrinsic functional properties of Parkin have not been addressed because of technical difficulties related to the development of a reliable in vitro model for the characterization of catalytic activity. We investigated the biochemical properties of Parkin and its pathogenic variants by analyzing in detail the effects of a series of 10 missense parkin gene mutations on the detergent solubility, tendency to form aggregates and protein–protein interactions of Parkin. Our results for this larger series of Parkin variants confirm our previous results, and those from other laboratories, showing that some of these proteins accumulate significantly in the detergent-insoluble cell fraction (30,32–35). This was the case for the five variants examined with amino-acid substitutions in the RING1 and RING2 fingers and for the ubiquitin-like domain variant, R42P. Four of these six variants (R275W, C289G, C418R, C441R) also tended to form visible aggregates when overproduced, although the lack of aggregation of the R42P and R256C variants suggests that decreasing Triton X-100 solubility is not necessarily correlated with aggregate formation. This observation is consistent with a previous attempt to categorize Parkin variants into ‘soluble’, ‘insoluble with low propensity to form inclusions’ and ‘insoluble with high propensity to form inclusions’ (33). The few discrepancies noted between our results and those obtained in this previous study may be due to differences in the cell types used. However, the sets of results available in the literature highlight the relevance of the RING and ubiquitin-like domains for the correct folding of Parkin. The integrity of these domains may also be important for protein stability, as suggested by (i) our previous comparative pulse-chase analysis showing that the C289G and C418G RING finger variants are cleared more rapidly than normal Parkin (30); (ii) a recent demonstration by pulse-chase analysis that mutations in the ubiquitin-like domain decrease Parkin stability (34); and (iii) the semi-quantitative analysis of relative amounts of protein provided in this study, in which we consistently observed lower total steady-state levels of the C418R, C441R and R42P variants. These observations are not consistent with parkin gene mutations compromising the ability of Parkin to promote its own degradation by autoubiquitylation, as previously suggested (38).
Surprisingly, none of the missense mutations studied abolished Parkin binding to a series of confirmed substrates (p38, CDCrel-1, -tubulin) and protein partners (Hsp70, -tubulin, proteasomal 4 subunit), as shown by GST pull-down analysis. The conservation of Parkin variant interactions with p38 is consistent with our previous observation of similar levels of p38-mediated sequestration for normal Parkin and the K161N, R256C, C289G and C418R variants in aggresomes in transfected cells (41), (C. Hampe and O. Corti, unpublished results). Coimmunoprecipitation experiments have also shown that the K161N, T240R and C431F parkin gene mutations do not affect the binding of Parkin to -tubulin (59) and that the interaction between Parkin and the proteasomal 4 subunit is preserved by the C289G and C418R mutations (60), (J. Dächsel, personal communication). However, although GST pull-down and coimmunoprecipitation techniques can be used to detect major interaction defects, they are generally not optimized for the reliable quantitative measurement of slight changes in protein binding capacity. We therefore cannot rule out the possibility that the mutations examined induce slight, but physiologically relevant changes in the affinity of Parkin for the tested protein interactors; alternatively, these changes may be undetectable in basal conditions, becoming significant under environmental stress. Previous studies have used coimmunoprecipitation or GST pull-down techniques to determine the effects of missense parkin gene mutations on the ability of Parkin to bind selected protein substrates/partners. In a few cases, the mutations were found to have strong effects. This was the case for the binding of Parkin to the E2 enzymes UbcH7 and UbcH8, which was abolished by the C-terminal T240R and T415N substitutions (26,38,61). Staropoli et al. also reported abrogation of the interaction of Parkin with the F-box protein hSel-10 by the RING1 T240R mutation, whereas Parkin binding to the chaperone-like protein 14-3-3
, was found to be abolished by the R42P, K161N and T240R mutations in another study (45,62). However, in other cases, mutations had little or no effect on the interactions between Parkin and selected proteins (32,42,43,48,59,60,63,64). In the absence of a functional test for determining the consequences of such modest effects, these results cannot be considered reliable and should be interpreted with caution. This is well illustrated by two recent studies reporting conflicting results concerning the effects of parkin gene mutations affecting the C-terminus of the protein on the interaction between Parkin and the substrate synphilin. The first of these studies suggested that RING2 mutations favor the binding of Parkin to synphilin (43), whereas the second postulated that RING2 mutations impair this interaction (32). This second study also reported various regulatory effects of a series of parkin gene mutations on the binding of Parkin to p38, an observation that is not supported by our data.
One of the major aims of our study was to develop a reconstituted purely in vitro autoubiquitylation assay for analysis of the intrinsic E3 ubiquitin–protein ligase activity of Parkin in an unbiased, substrate- and cofactor-independent manner. Indeed, in vitro assays of the autoubiquitylation of proteins with E3 ubiquitin–protein ligase activity is a valuable, recognized means of assaying the catalytic activity of these proteins (65–68). The development of this test was hampered by the marked tendency of bacterially produced GST–Parkin to be sequestered in inclusion bodies. Despite this difficulty, small amounts of soluble GST–Parkin were purified and displayed autoubiquitylation activity in vitro. This test demonstrated unambiguously that most of the missense parkin gene mutations tested did not abolish Parkin enzymatic activity. Only two of 10 pathogenic variants tested were inactive; both mutations concerned highly conserved cysteine residues of RING2. Neither of these variants affected the catalytic activity of normal Parkin on coincubation, suggesting that inactive pathogenic Parkin variants do not exert dominant negative effects on active variants. Moreover, none of the eight additional variants examined showed reproducible differences in catalytic activity with respect to normal Parkin. During the preparation of this manuscript, Matsuda et al. (57) reported similar results using an in vitro ubiquitylation test based on a maltose binding protein (MBP)–Parkin fusion protein produced in bacteria. These findings and our study demonstrate that only mutations replacing essential amino acids in the RING2 domain or truncating this domain, abolish Parkin enzymatic activity. These results strongly indicate that RING2 plays an essential role in the catalytic core of the protein, as suggested in a previous report identifying the RING2 domain of HHARI, a member of the RING-IBR-RING/TRIAD protein family, as the domain responsible for the E3 ubiquitin–protein ligase activity of the protein (69).
These results clarify the conflicting results obtained in previous attempts to explore the consequences of parkin gene mutations on the E3 ubiquitin–protein ligase activity of Parkin. Some of these studies were based on analysis of the ubiquitylation state of Parkin and its pathogenic variants in transfected cells (32,38,45,46). Our results demonstrate that this test is not a reliable indicator of the intrinsic catalytic activity of Parkin. Indeed, whereas these previous studies have attributed differences in the ubiquitylation pattern of Parkin proteins overproduced in cells to differences in Parkin autoubiquitylation, we did not detect reliable differences in the autoubiquitylation potential of equal amounts of a series of active pathogenic Parkin variants. In particular, parkin gene mutations that were clearly inactive, based on our own results and those of Matsuda et al. (i.e. the C418R and C431F missense and the W453X truncating mutations), have been shown to be ubiquitylated in cell models (30,32). Therefore, inactivating gene mutations do not preclude ubiquitylation of the corresponding proteins by other ligases in cells. Evidence that this may be the case was provided by the recent identification of an E3 ubiquitin–protein ligase involved in Parkin degradation (70).
In another series of studies, the enzymatic activity of pathogenic Parkin variants was investigated by (i) exploring the ability of overproduced Parkin to promote the ubiquitylation of a selected substrate in transfected cells, or (ii) analyzing semi-in vitro the ubiquitylation of substrates by a Parkin protein immunoprecipitated from cell lysates or generated by translation in reticulocyte lysates (32,39–48). Both approaches are potentially biased by their substrate dependence, as the loss of an interaction with a particular substrate inevitably results in a lack of ubiquitylation, irrespective of the intrinsic catalytic activity of Parkin. Endogenous E3 ubiquitin–protein ligases, modulators of Parkin activity or other cofactors of ubiquitylation reactions, the abundance of which may depend on the cell type used, may also act as sources of ambiguity, potentially accounting for the high background levels of ubiquitylation often observed in the absence of overproduced Parkin (32,44,46). Because of these biases, contradictory results have generally been obtained with these experimental approaches: e.g. we observed that the K161N, R256C and C289G variants ubiquitylated the p38 substrate with similar efficiencies when overproduced in COS7 cells (41) (C. Hampe and O. Corti, unpublished results), whereas other studies found these variants to be inactive or significantly less active than normal Parkin (32,42–44,46). Similarly, Shimura et al. reported the R42P variant to be inactive, whereas Sriram et al. and Ko et al. concluded that it promoted the ubiquitylation of synphilin or p38 even more efficiently than normal Parkin (26,32,40,48). Finally, whereas Chung et al. observed that the T415N and Q311X mutations preserved the ability of Parkin to ubiquitylate synphilin in HEK293 cells, Sriram et al. reported that these mutations strongly decreased the ubiquitylation of synphilin in SH-SY5Y cells (32,43).
In addition to clarifying the effect of missense parkin gene mutations on the enzymatic activity of the protein, the development of a Parkin-dependent in vitro autoubiquitylation assay enabled us to analyze the type of ubiquitylation promoted intrinsically by Parkin. Using a lysine-less ubiquitin derivative and antibodies specifically recognizing polyubiquitylated proteins rather than both mono- and polyubiquitylated proteins, we demonstrated that, unlike the ubiquitin–protein ligases, Mdm2 and CHIP, Parkin promoted the attachment of single ubiquitin molecules only. Matsuda et al. recently came to a similar conclusion, using MBP–Parkin and methylated or lysine-less ubiquitin (57). These results suggest that Parkin may play a role in the monoubiquitylation of proteins in vivo, which is a reversible non-proteolytic post-translational modification involved in various cellular processes, including the endocytic internalization of membrane receptors, the regulation of histone activity, bacterial entry into cells and virus budding (49,50,71). The role of Parkin in cellular protein monoubiquitylation has not been investigated, although a pattern of Parkin-dependent ubiquitylation compatible with the conjugation of a single ubiquitin molecule was reported in a previous study (47). In contrast, previous studies have addressed the type of ubiquitin-modification promoted by Parkin in polyubiquitylated proteins. In particular, Lim et al. showed that Parkin mediates the polyubiquitylation of synphilin via both Lys-48- and Lys-63-linked ubiquitin chains, and suggested a role for this modification in the formation of synphilin-positive LB-like inclusions (55). Similarly, Doss-Pepe et al. (56) observed enhanced ubiquitin conjugation in a rabbit reticulocyte fraction upon the addition of bacterially produced Parkin together with ubiquitin derivatives containing a single Lys-48 or Lys-63 residue. However, this study did not analyze, in parallel, the effects of a lysine-less ubiquitin derivative in this model. We therefore cannot exclude the possibility that the monoubiquitylation of proteins at single or multiple lysines accounts for at least some of the ubiquitin conjugation observed. Further studies are required to determine the extent to which Parkin catalyzes the attachment of single ubiquitin molecules in vivo, but both our results and those of these previous reports suggest that Parkin may promote monoubiquitylation or different types of polyubiquitylation, depending on the cellular environment and protein context. According to this hypothesis, protein monoubiquitylation is promoted by Parkin alone, whereas the proteolytic and non-proteolytic conjugation of polyubiquitin chains is mediated by Parkin in association with cellular factors. The protein chaperones, Hsp70, 14-3-3 and BAG5, and the U-box-dependent E3 ligase, CHIP, have been shown to associate with Parkin and to modulate its activity (47,62,63,72). Parkin also functions in a multiprotein Skp1-, Cullin- and F-box (SCF)-like complex displaying E3 ubiquitin–protein ligase activity enhanced by the F-box protein, hSel-10 (45). Future studies are required to determine whether these or other as yet unknown regulatory interactions induce a switch in the type of ubiquitin conjugation promoted by Parkin. We found that Parkin is modified by multimonoubiquitylation in cells, although it is unclear whether this modification is mediated by Parkin itself or induced by other E3 ubiquitin–protein ligases. It is tempting to speculate that this post-translational modification of Parkin regulates the interactions of this protein with other cellular proteins, its subcellular location and/or enzymatic activity, thereby having a direct or indirect impact on the type of ubiquitylation that it promotes.
In conclusion, our results demonstrate that the loss of intrinsic E3 ubiquitin–protein ligase activity is a rare consequence of missense parkin gene mutations. In contrast, consistent with previous studies, a decrease in the detergent solubility of Parkin has emerged as the most common effect of these mutations: in this study, more than half of the amino-acid substitutions examined promoted the accumulation of Parkin in the Triton X-100-insoluble cell fraction and some of these substitutions also appeared to decrease protein stability. Detergent insolubility may reflect protein mislocalization rather than misfolding, as it is not correlated with a loss of enzymatic activity. It may lead to the depletion of Parkin from essential sites of action and result in a loss of function phenotype similar to that induced by inactivating missense or truncating mutations. The mechanisms underlying the pathogenic effects of a series of missense parkin gene mutations that seem to preserve the stability, subcellular distribution, protein interactions and enzymatic activity of Parkin remain to be elucidated. One of these mechanisms could be the predisposition to stress-induced solubility alterations, as suggested by a recent study showing that two soluble Parkin variants are more prone to rotenone-induced changes in solubility when compared with normal Parkin (36). However, it is also possible that some of the missense point mutations reported in the literature are in fact polymorphisms without pathogenic consequences. This may well be the case of the A82E substitution, which affects an amino-acid position that is poorly conserved throughout evolution and which was found only in the heterozygous state, or in the homozygous state, but with an associated homozygous parkin gene deletion (Table 1). Increasing our understanding of the dynamic regulation of Parkin distribution within the cell, of the network of Parkin intermolecular interactions, and of the cooperation between these parameters in modulating the ubiquitylation capacity of this protein, should help better understand the functional consequence of parkin gene mutations. This would also constitute an essential step towards the full comprehension of the potentially multiple physiological functions of Parkin and their relationship with disease.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Plasmids, cell culture and transfection
The Parkin cDNA was inserted between the EcoRI and XbaI sites of pcDNA3-HA (A. Fournier). Parkin mutants were obtained by site-directed mutagenesis (Stratagene). For bacterial expression, the HA-Parkin cDNAs were inserted between the BamHI and NotI sites of pGEX-6P1 (Amersham Biosciences). The p38, CDCrel-1 and
-synuclein cDNAs were inserted between the BamHI and NotI, EcoRI and XhoI and XhoI and XbaI sites of pcDNA3-myc (L. Pradier), respectively. The plasmids for the eukaryotic expression of HA-ubiquitin (pMT123) or normal and lysine-less His-ubiquitin (pCB6-6H-Ub2-WT/K0 were kindly provided by G. Bossis and R. Baer. The integrity of the constructs was confirmed by sequencing. COS7 cells were grown in DMEM (Invitrogen) supplemented with 10% fetal calf serum in a 5% CO2 atmosphere. Cells were transfected using DMRIE-C (Invitrogen) according to the manufacturer's instructions.
Preparation of cell fractions and anti-HA-Parkin immune complexes
COS7 cells were transfected with pcDNA3-HA encoding wild-type or mutant Parkin proteins. Forty-eight hours after transfection, cells were harvested in phosphate-buffered saline (PBS) and resuspended in lysis buffer [50 mM Tris–HCl (pH 8.0), 300 mM NaCl, 1.5 mM MgCl2, 1% Triton X-100, complete protease inhibitors (Roche)]. The whole cell extract (corresponding to the total fraction, T) was separated by centrifugation (15 000g, 30 min, 4°C) to obtain the Triton X-100-soluble (S) and insoluble fractions (pellets, P). The pellets were treated with denaturing lysis buffer (6 M guanidium–HCl, 0.1 M Na2HPO4/NaH2PO4, 0.01 M Tris–HCl, pH 8.0) and proteins were precipitated with 5% trichloroacetic acid (TCA). We analyzed 30 µg of the total and the soluble fraction and 1/40 of the pellet by western blotting with mouse monoclonal anti-HA-antibodies (HA.11, clone 16B12, 1:2000, BabCO) or anti-Parkin antibodies (clone PRK8, 1:12000, Upstate). The membrane was then probed with rabbit polyclonal anti-actin antibody (A2066, 1:2000, Sigma). The signals were quantified by densitometry using Mercator image analysis software, Proversion V2.20 (Explora Nova).
To obtain anti-HA-Parkin immune complexes, COS7 cells overproducing HA-Parkin or HA-Parkin variants were lysed in lysis buffer [20 mM HEPES (pH 7.9), 150 mM KCl, 1% Triton X-100, 1.5 mM MgCl2, 0.5% glycerol, 0.2 mM Na3VO4, 4 mg/ml NaF, 5.4 mg/ml ß-glycerophosphate, complete protease inhibitors (Roche)] 48 h after transfection. Lysates cleared by centrifugation were incubated with monoclonal anti-HA antibodies (HA.11, clone 16B12, BabCO) for 2 h, then with protein G-sepharose beads for 1 h. Immune complexes were washed either five times with lysis buffer and four times with ubiquitylation buffer [50 mM Tris–HCl (pH 7.2), 120 mM NaCl, 5 mM MgCl2; stringent conditions], or five times with ubiquitylation buffer (less stringent conditions).
Immunocytochemistry
COS7 cells were transfected with pcDNA3-HA encoding wild-type or mutant Parkin. After 48 h, the cells were fixed in 4% paraformaldehyde and analyzed by standard immunocytochemical procedures, with rabbit polyclonal anti-Parkin (#2132, 1:2000, Cell Signaling) and Cy3-conjugated goat anti-rabbit IgG (Jackson Immunoresearch, 1:2000).
Production of recombinant proteins
GST–Parkin mutants were produced in Rosetta (DE3)pLys cells (Novagen) in the exponential growth phase. Protein production was induced by adding 0.1 mM isopropyl- 1-thio-ß-D-galactopyranoside (IPTG) at 30°C for 2 h. GST–Mdm2 was produced as described elsewhere (67). Bacterial pellets were resuspended in PBS, incubated on ice for 30 min and lysed by sonication. Triton X-100 was added to a final concentration of 1% and the sonicate was then incubated on ice and clarified by centrifugation at 4°C for 40 min at 29 000g. The recombinant proteins were batch purified with glutathione-sepharose (Amersham Biosciences) according to the manufacturer's instructions. For in vitro ubiquitylation assays, the GST-fusion proteins were eluted from the beads by adding 15 mM reduced glutathione (Sigma). Proteins were dialyzed, quantified, aliquoted and stored at –80°C. Where indicated, glutathione-sepharose-bound GST-HA-Parkin was digested with PreScission protease (Amersham Biosciences). Before digestion, the beads were washed three times with 1% Triton X-100 in PBS, and once with cleavage buffer [50 mM Tris–HCl (pH 7.2), 120 mM NaCl, 5 mM MgCl2, 1 mM DTT]. GST-HA-Parkin (20 µg) was cleaved on the beads by incubation with 40 units PreScission protease for 4 h at 5°C. The GST-moiety and the protease were then removed by immobilization on fresh glutathione-sepharose beads for 20 min at 4°C. The recombinant proteins were separated by SDS-PAGE and quantified on gels stained with Coomassie Blue.
GST pull-down assays
COS7 cells were transfected with pcDNA3-myc--synuclein, pcDNA3-myc-p38 or pcDNA3-myc-CDCrel-1. After 48 h, cells were harvested in PBS and resuspended in pull-down buffer [20 mM HEPES (pH 7.9), 150 mM KCl, 1% Triton X-100, 1.5 mM MgCl2, 0.5 mM DTT, 0.2 mM Na3VO4, 4 mg/ml NaF, 5.4 mg/ml ß-glycerophosphate, complete protease inhibitors (Roche)]. The resulting lysates were centrifuged (15 000g, 30 min, 4°C), and the supernatants precleared by incubation for 3 h at 4°C with 15 µg GST per milligram of lysate. We incubated 5 µg of GST or GST–Parkin mutants immobilized on glutathione-sepharose beads (Amersham Biosciences) with 500 µg of cleared COS7 cell lysate overnight at 4°C. Beads were washed four times in pull-down buffer and boiled in protein sample buffer [250 mM Tris–HCl (pH 6.8), 500 mM DTT, 10% SDS, 0.5% bromophenol blue, 50% glycerol]. The released proteins were resolved by SDS–PAGE, and analyzed by western blotting with the following mouse monoclonal antibodies: anti-myc (clone 9E10, 1:600, Santa Cruz), anti--tubulin (clone GTU-88, 1:10000, Sigma), anti--tubulin (clone DM 1A, 1:1000, Sigma), anti-Hsp70 (clone C92F3A-5, 1:1000, Stressgen), or anti-20S proteasome subunit 4 (clone MCP34, 1:1000, Biomol).
In vitro ubiquitylation assays
Standard ubiquitylation assays were carried out for 90 min at 30°C, in a final volume of 40 µl of ubiquitylation buffer [50 mM Tris–HCl (pH 7.2), 120 mM NaCl, 5 mM MgCl2] containing 0.5 mM DTT, and 4 mM ATP. For in vitro ubiquitylation, normal or mutated GST-HA-Parkin (1 µg), or anti-HA-Parkin immune complexes, were incubated in the presence of 100 ng E1 (Sigma), 500 ng E2 UbcH7 (Sigma) and 2 µg FLAG-ubiquitin (Sigma). If GST–Mdm2 or HA–CHIP were used as the E3 ubiquitin–protein ligase, UbcH5B (Biomol) or UbcH5C (BostonBiochem) were used as the E2 ubiquitin-conjugating enzyme. Where indicated, 3 µg of His-tagged wild-type or lysine-less (K0) ubiquitin (BostonBiochem) were used instead of FLAG-ubiquitin. Once the reaction was complete, the GST-fusion proteins were purified by incubation with glutathione-sepharose beads (Amersham Biosciences) for at least 3 h at 4°C. The beads were then washed four times in washing buffer [50 mM Tris–HCl (pH 8.0), 500 mM NaCl, 1.5 mM MgCl2, 1% Triton X-100]. Supernatants and beads were boiled in protein sample buffer and resolved by SDS–PAGE on 3–8% (Parkin, Mdm2) or 4–12% (CHIP) gradient gels (Invitrogen). Gels were electroblotted onto a nitrocellulose membrane and probed with rabbit polyclonal anti-ubiquitin antibodies (Z 0458, 1:1000, DakoCytomation), mouse monoclonal anti-FLAG antibodies (M2, 1:10000, Sigma), mouse monoclonal antibodies against polyubiquitylated proteins (clone FK1, 1:1000, Biomol) or antibodies specific for mono- and polyubiquitylated proteins (clone FK2, 1:2000, Biomol).
Ubiquitylation of Parkin and p38 in COS7 cells
To analyze the pattern of ubiquitylation of Parkin with FK1 and FK2 antibodies, HA-tagged ubiquitin was overproduced alone or with FLAG-Parkin in COS7 cells. Where indicated, the cells were treated for 6 h with epoxomicin (1 µM) before harvesting. Forty-eight hours after transfection, the cells were harvested in PBS and resuspended in lysis buffer [20 mM HEPES (pH 7.9), 150 mM KCl, 1% Triton X-100, 1.5 mM MgCl2, 0.5% glycerol, 0.2 mM Na3VO4, 4 mg/ml NaF, 5.4 mg/ml ß-glycerophosphate, 20 mM N-ethylmaleimide (Sigma), complete protease inhibitors (Roche)]. The lysate was cleared by centrifugation and FLAG-Parkin immunoprecipitated overnight at 4°C in the presence of rabbit polyclonal anti-Parkin antibody (#2132, Cell Signaling) conjugated to protein G-sepharose beads (Amersham Biosciences). Beads were washed four times in washing buffer [20 mM HEPES (pH 7.9), 500 mM KCl, 1.5 mM MgCl2, 1.5% Triton X-100, 0.2 mM Na3VO4, 4 mg/ml NaF, 5.4 mg/ml ß-glycerophosphate, 20 mM N-ethylmaleimide, complete protease inhibitors], boiled in protein sample buffer and the released proteins were resolved by SDS–PAGE (3–8% gradient gel, Invitrogen). The ubiquitylation of immunoprecipitated FLAG-Parkin was analyzed by western blotting with the following mouse monoclonal antibodies: FK1 (1:1000, Biomol), FK2 (1:2000, Biomol), and anti-HA (HA.11, clone 16B12, 1:2000, BabCO). The membranes were then probed with mouse monoclonal anti-FLAG M2 (1:10 000, Sigma).
The patterns of ubiquitylation of Parkin and p38 obtained in the presence of normal or lysine-less ubiquitin were analyzed in COS7 cells overexpressing FLAG-Parkin or HA-Parkin and myc-p38, together with normal or lysine-less His-tagged ubiquitin. The cells were lysed in denaturing conditions (6 M guanidium–HCl, 0.1 M Na2HPO4/NaH2PO4, 0.01 M Tris–HCl, pH 8.0, 500 mM NaCl) and the His-tagged ubiquitylated proteins were purified as previously described (41). Ubiquitylated proteins and aliquots of the total cell fractions precipitated with 5% TCA were analyzed by western blotting with anti-FLAG M2 (1:10000, Sigma) and anti-myc antibodies (clone 9E10, 1:600, Santa Cruz).
|
ACKNOWLEDGEMENTS |
|---|
We thank S. Lesage and C. Depienne for helpful discussions, A.M. Weissman for kindly providing pGEX-4T1-Mdm2, W.S. Trimble for providing pGEX-KG-CDCrel-1, Y. Imai and R. Takahashi for the expression plasmids encoding FLAG-Parkin and HA–CHIP, M. Ghee and J. Mallet for the human
-synuclein cDNA, G. Bossis for the pHA-ubiquitin vector (pMT123) and R. Baer for the pCB6-6H-Ub2-WT/K0 vectors. This work was supported by INSERM, the Fondation de France, and APOPIS (Abnormal proteins in the pathogenesis of neurodegenerative disorders—an integrated project funded by the EU under the Sixth Framework Programme; Priority: Life Science for Health, contract no. LSHM-CT-2003-503330). C. Hampe was supported by a fellowship from the Association France Parkinson. Conflict of Interest statement. None declared.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
- de Rijk M.C., Tzourio C., Breteler M.M., Dartigues J.F., Amaducci L., Lopez-Pousa S., Manubens-Bertran J.M., Alperovitch A., Rocca W.A. (1997) Prevalence of parkinsonism and Parkinson's disease in Europe: the EUROPARKINSON Collaborative Study. European Community Concerted Action on the Epidemiology of Parkinson's disease. J. Neurol. Neurosurg. Psychiat. 62:10–15.[Abstract]
- Forno L.S. (1996) Neuropathology of Parkinson's disease. J. Neuropathol. Exp. Neurol. 55:259–272.[ISI][Medline]
- Gasser T. (2005) Genetics of Parkinson's disease. Curr. Opin. Neurol. 18:363–369.[ISI][Medline]
- Zimprich A., Biskup S., Leitner P., Lichtner P., Farrer M., Lincoln S., Kachergus J., Hulihan M., Uitti R.J., Calne D.B., et al. (2004) Mutations in LRRK2 cause autosomal-dominant parkinsonism with pleomorphic pathology. Neuron 44:601–607.[CrossRef][ISI][Medline]
- Paisan-Ruiz C., Jain S., Evans E.W., Gilks W.P., Simon J., Van Der B.M., Lopez D.M., Aparicio S., Gil A.M., Khan N., et al. (2004) Cloning of the gene containing mutations that cause PARK8-linked Parkinson's disease. Neuron 44:595–600.[CrossRef][ISI][Medline]
- Di Fonzo A., Rohe C.F., Ferreira J., Chien H.F., Vacca L., Stocchi F., Guedes L., Fabrizio E., Manfredi M., Vanacore N., et al. (2005) A frequent LRRK2 gene mutation associated with autosomal dominant Parkinson's disease. Lancet 365:412–415.[ISI][Medline]
- Nichols W.C., Pankratz N., Hernandez D., Paisan-Ruiz C., Jain S., Halter C.A., Michaels V.E., Reed T., Rudolph A., Shults C.W., et al. (2005) Genetic screening for a single common LRRK2 mutation in familial Parkinson's disease. Lancet 365:410–412.[ISI][Medline]
- Kachergus J., Mata I.F., Hulihan M., Taylor J.P., Lincoln S., Aasly J., Gibson J.M., Ross O.A., Lynch T., Wiley J., et al. (2005) Identification of a novel LRRK2 mutation linked to autosomal dominant parkinsonism: evidence of a common founder across European populations. Am. J. Hum. Genet. 76:672&