Dorfin Localizes to Lewy Bodies and Ubiquitylates Synphilin-1

Parkinson’s disease (PD) is a neurodegenerative disease characterized by loss of nigral dopaminergic neurons. Lewy bodies (LBs) are a characteristic neuronal inclusion in PD brains. In this paper, we report that Dorfin, a RING-finger type ubiquityl ligase for mutant SOD1, was localized with ubiquitin in LBs. Recently, synphilin-1 was identified to associate with a -synuclein and to be a major component of LBs. We found that overexpression of synphilin-1 in cultured cells led to the formation of large juxtanuclear inclusions, but showed no cytotoxicity. Dorfin colocalized in these large inclusions with ubiquitin and proteasomal components. In contrast to full-length synphilin-1, overexpression of the central portion of synphilin-1 including ankyrin-like repeats, a coiled-coil domain and an ATP, GTP-binding domain, predominantly led to the formation of small punctate aggregates scattered throughout the cytoplasm and showed cytotoxic effects. Dorfin and ubiquitin did not localize in these small aggregates. Overexpression of the N- or C-terminus of synphilin-1 did not lead to the formation of any aggregates. Dorfin physically bound and ubiquitylated synphilin-1 through the central portion of synphilin-1, but ubiquitylated neither wild-type nor mutant a -synuclein. These results suggest that the central domain of synphilin-1 has an important role in the formation of aggregates and cytotoxicity, and that Dorfin may be involved in the pathogenetic process of PD and LB formation by ubiquitylation of synphilin-1. our (ii) parallel distribution patterns of ubiquitin and Dorfin and inclusion bodies synphilin-1 in and (iii) the E3 of synphilin-1. our


Introduction
Parkinson's disease (PD) 1 is a neurodegenerative disease caused by loss of nigral dopaminergic neurons. The Lewy body (LB) is a characteristic neuronal inclusion in brains with PD (1)(2)(3)(4). Although LBs are a prominent pathologic feature of PD, the underlying molecular mechanism accounting for LB formation is poorly understood.
Several lines of evidence have suggested that derangements in the ubiquitinproteasomal protein degradation pathway play a prominent role in the pathogenesis of PD (5). Ubiquitin and proteasome subunits colocalize in LB (6,7), and biochemical studies have revealed reduced catalytic activities of proteasomes in the lesions of PD (8,9). The gene product responsible for autosomal recessive juvenile parkinsonism (AR-JP), parkin (10), is an E3 ubiquityl ligase (11)(12)(13). Accumulation of target protein(s) due to loss of the ubiquitylation function of parkin may contribute to the development of AR-JP. In addition, a missense mutation in ubiquitin C-terminal hydrolase L1 (UCHL1) has been described in a family with PD (14). UCHL1 produces monomeric ubiquitin by cleaving polyubiquitin chains (15). Recently, ubiquityl ligase activity as well as the hydrolase activity of UCHL1 were also reported (16).
a-Synuclein is a 19-kDa presynaptic vesicular protein of unconfirmed function and one of the major components of LBs (17,18). Mutations in a-synuclein (A30P and A53T) cause a rare autosomal dominant form of PD, which share many phenotypic findings with sporadic PD (19,20). a-Synuclein aggregates deposit in LBs by guest on March 22, 2020 http://www.jbc.org/ Downloaded from Ito et al., Page 4 in both autosomal dominant and sporadic PD (21,22). In addition, it has been reported that transgenic flies and mice overexpressing human wild-type or mutant a-synuclein have abnormal cellular accumulation of a-synuclein and neuronal dysfunction and degeneration (23)(24)(25)(26)(27)(28)(29)(30), indicating that a-synuclein has a role in the pathogenesis of both familial and sporadic PD.
Synphilin-1 was identified recently by yeast two-hybrid techniques to be a novel protein interacting with a-synuclein (31). a-Synuclein amino acid 1-65 region is sufficient for interaction, and the central portion of synphilin-1 (amino acids 349-555) is necessary and sufficient for interaction with a-synuclein (32). It has also been reported that the C-terminus of a-synuclein is closely associated with the C-terminus of synphilin-1 and a weak interaction occurs between the N-terminus of a-synuclein and synphilin-1 (33). Synphilin-1 is highly concentrated ed in presynaptic nerve terminals and its association with synaptic vesicles is modulated by a-synuclein (34).
Coexpression of a-synuclein and synphilin-1 in transfected cells results in the formation of eosinophilic cytoplasmic inclusions that resemble LBs (31,35), while transfection of synphilin-1 alone without expression of a-synuclein or parkin can also produce cytoplasmic inclusions in cultured cells (36,37). Furthermore, synphilin-1 is ubiquitylated and degraded by proteasomes in human embryonic kidney 293 (HEK293) cells (37), and is localized as another major component of LB in the brains of patients with PD (38,39). Thus, the process through which aggregations are formed by synphilin-1 may be important in the pathogenesis of PD. Dorfin is a gene product, which we cloned from anterior horn tissues of the human spinal cord (40), that contains a RING-finger/IBR motif (41) at its N-terminus.
It was reported that HHARI (human homologue of ariadne) and H7-AP1 (UbcH7associated protein), both RING-finger/IBR motif-containing proteins, interact with ubiquitin-conjugating enzyme (E2) UbcH7 through the RING-finger/IBR motif, and that a distinct subclass of RING-finger/IBR motif-containing proteins represents a new family of proteins that specifically interact with distinct E2 enzymes (42,43).
Dorfin is a juxtanuclearly located E3 ubiquityl ligase, and may function in the microtubule organizing centers (MTOCs) (40). In the spinal cords of patients with sporadic and familial amyotrophic lateral sclerosis (ALS) with SOD1 mutation, Dorfin is colocalized with ubiquitin in hyaline inclusions (44). Dorfin physically bound and ubiquitylated various SOD1 mutants derived from familial ALS patients and enhanced their degradation (44). Thus, an important and interesting question is whether Dorfin is colocalized with ubiquitin in LBs of PD.
In this study, we showed that Dorfin is colocalized with ubiquitin in LBs of PD.
We found that Dorfin ubiquitylates synphilin-1 and that overexpression of synphilin-1 leads to ubiquitylated inclusions resembling LBs in cultured cells.

Experimental procedures
Immunohistochemistry --Immunohistochemical studies were carried out on 20% buffered formalin-fixed, paraffin-embedded autopsied brains filed in the Department of Neurology, Nagoya University Graduate School of Medicine. Five PD brains (age: 67-69 years, four men and one woman) and five controls without neurological disease (age: 61-78 years, four men and one woman) were studied. The diagnosis of all cases was confirmed by clinical and pathological criteria. Immunohistochemistry was performed as described previously (45). Polyclonal rabbit antiserum was raised against a C-terminal portion of Dorfin, amino acid 678-690, as described previously (40). Dorfin antiserum (l:200 dilution) and monoclonal anti-ubiquitin antibody (P4D1, 1:400 dilution; Santa Cruz Biotechnology) were used. To assess the colocalization of Dorfin and ubiquitin, a double-labelling immunofluorescence study was performed on

Dorfin Localizes to Lewy Bodies of PD --We first examined whether LBs contain
Dorfin. Immunohistochemical analysis revealed that Dorfin was predominantly localized in LBs found in PD (Fig. 1A). The peripheral rim of a typical LB in a neuronal cell body was strongly stained, while the central core remained unstained of Dorfin was very similar to that of a-synuclein (48), which is predominantly located in the peripheral rim of LBs, but was different from that of parkin, which localizes predominantly in the core of LBs (49).

Inclusions as Full-length Proteins, but Small Punctate Aggregates Are Also
Formed --To further analyze which part of synphilin-1 is related to aggregation formation, we prepared a series of deletion mutants of synphilin-1. We divided  (Fig. 3L-Q).  (37). Thus, we examined the effects of synphilin-1 or its deletion mutants on cell viability using MTS assay in a neuronal cell line, Neuro2a

Expression of the Central Portion of Synphilin-1 Compromises Cell Viability
( Fig. 4C). We found that synphilin-1-M had a cytotoxic effect, whereas overexpression of full-length synphilin-1, or N-terminal or C-terminal deletion mutants of synphilin-1, did not (Fig. 4C). We used synphilin-1-V5 fusion constructs, but synphilin-1-DsRed fusion constructs gave the same results (data not shown). Dorfin in COS7 cells (Fig. 5A). Co-immunoprecipitation confirmed that Dorfin bound to full-length synphilin-1 (Fig. 5B) and interacted with synphilin-1-M strongly and synphilin-1-N weakly, but Dorfin failed to bind synphilin-1-C (Fig. 5C). Thus, Dorfin interacts with synphilin-1 mainly through its central portion, which contains the ankyrin-like repeat, the coiled-coil domain and the ATP/GTP binding domain. Dorfin has a unique primary structure containing a RING-finger/IBR motif at its N-terminus and can be structurally divided into two parts, the N-terminal region containing a RING-finger/IBR motif (Dorfin-N) that interacts with E2, and the C-terminal region with no similarity to any other known proteins (Dorfin-C) (40) (Fig. 5A). We found that Dorfin-C, but not Dorfin-N, specifically bound synphilin-1, indicating that Dorfin binds to synphilin-1 via its C-terminal region (Fig. 5D).
We next examined whether Dorfin is involved in the ubiquitylation of synphilin-1 in vitro. For this purpose, we immunopurified Xpress-Dorfin and synphilin-1-V5 independently after transfection into HEK293 cells. When these immunopurified proteins were incubated with recombinant E1, E2 (UbcH7), Histagged ubiquitin, and ATP, high molecular weight ubiquitylated bands were observed in the presence of Xpress-Dorfin with synphilin-1, while no signal was noted in synphilin-1 in the absence of either E1 or E2 (Fig. 6B). Dorfin ubiquitylated mutant SOD1 in vitro, as previously reported (44). Dorfin ubiquitylated neither wild-type nor mutant a-synuclein (Fig. 6B). In vitro ubiquitylation assay of a series of synphilin-1 deletion mutants by Dorfin revealed synphilin-1-M was ubiquitylated, whereas synphilin-1-N or synphilin-1-C were not ubiquitylated at all (Fig. 6C). whereas the small aggregates contained non-fibrillar spherical aggregates (50). They suggested that these aggregates appear sequentially, with the smallest population appearing the earliest and the fibrillar inclusions the latest, and that the small spherical aggregates are the cellular equivalents of protofibrils (50). Protofibrils are recognized to be more important in the terms of cytotoxicity than mature fibrils in Ab (51,52) and a-synuclein (53,54). In our cell culture model, overexpression of synphilin-1 produced two distinct types of aggregates, very closely resembling two types of asynuclein aggregates (50). Thus, the small punctate aggregates scattered throughout the cytoplasm induced by the central portion of synphilin-1 might have characteristic similar to those of protofibrils. Our cell culture system will allow detailed characterization of LB formation and cytotoxic processes in further studies.
We reported previously that Dorfin localizes in the inclusion bodies of familial ALS with SOD1 mutations as well as in those of sporadic ALSs, and ubiquitylates various SOD1 mutants derived from familial ALS patients (44). Based on these findings, it is conceivable that familial and sporadic forms of ALS share a common Parkin has been shown to have E3 ubiquityl ligase activity (10)(11)(12). It was recently demonstrated that an O-glycosylated a-synuclein (55) and synphilin-1 (35) are the substrates of parkin, and that parkin localizes to LBs of sporadic PD (49). The link between sporadic and familial PD through a-synuclein, synphilin-1 and parkin sheds new light on underlying common molecular pathogenetic mechanisms in PD.
What roles, then, do Dorfin and parkin play with respect to each other in the