DNAJC proteins and pathways to parkinsonism

Abstract

The DNAJC protein family is a subclass of heat shock proteins that has attracted recent attention due to the identification of mutations that are linked with parkinsonism, a feature of Parkinson’s disease and other neurological disorders. In this review we discuss the current genetic and functional evidence of the association of these DNAJC proteins with disease and how mutations in these proteins may contribute to disease pathogenesis. Whereas DNAJC6 (Auxilin), DNAJC12 and DNAJC5 (CSPα) exhibit strong genetic association with disease, DNAJC26 (GAK), DNAJC13 (RME-8) and DNAJC10 (Erdj5) require additional evidence to definitively link reported variants to parkinsonism. Remarkably, multiple DNAJC proteins (Auxilin, GAK, RME-8, CSPα) functionally converge on pathways of synaptic trafficking and clathrin dynamics, highlighting an important role of those pathways in the pathogenesis of parkinsonism. Further research is required to define the mechanisms through which these mutations contribute to disease etiology.

Graphical Abstract

A number of DNAJC proteins have been recently linked to risk of Parkinson’s disease, uncovering a potential new pathway resulting in neurodegeneration. In this review, the evidence linking DNAJC proteins to parkinsonism is summarised, with potential mechanisms connecting these proteins to neuronal cell death discussed. (Images adapted from An Essay on the Shaking Palsy, J. Parkinson, 1817, and Manual of the Diseases of the Nervous System, W. Gower, 1886; clathrin cage derived from PDB 1xi5).

1. Introduction

Parkinson’s disease (PD) is a common neurodegenerative disorder and is characterized by a range of motor symptoms including resting tremor, postural instability, muscle rigidity and bradykinesia - collectively described as parkinsonism - as well as by an increasingly appreciated spectrum of non-motor symptoms. These Parkinsonian symptoms are driven by the degeneration of dopaminergic neurons in the substantia nigra pars compacta, accompanied by the presence of Lewy pathology in surviving neurons [1]. Although parkinsonism is a central clinical feature of PD, it also found in other neurological diseases that are grouped as secondary parkinsonisms (that is, the presence of Parkinsonian symptoms, but deriving from distinct disease etiologies leading to dysfunction of the dopaminergic system), or atypical parkinsonism, which present additional clinical features not found in idiopathic PD, such as ataxia and seizures [2,3]. As there is significant overlap in the clinical presentation as well as genetic background of the spectrum of typical and atypical parkinsonisms, it has been proposed that the underlying disease mechanisms could converge on common pathological pathways [2,3]. In this context, members of the DnaJ homolog C (DNAJC) family, a subclass of the heat shock protein (HSP) family, have attracted recent attention as mutations in multiple members of this protein family have been associated with PD and other parkinsonisms (Table 1).

Table 1:

Overview of clinical features of DNAJ proteins associated with PD and parkinsonisms.

Gene	Protein	Function	Mutations	Mode of inheritance	Age of onset	Clinical presentation	References
DNAJC6	Auxilin 1	Uncoating of clathrin coated vesicles	c.802–2A>G, p.T741=, p.Q791*, p.Q846*, p.R927G	Autosomal recessive	8–45	Juvenile and early onset PD. Cardinal motor symptoms with additional neurological features including epilepsies, cognitive decline, psychiatric and pyramidal signs, limited L-DOPA response due to side effects	[23–26]
DNAJC26	Cyclin G-dependent kinase A (GAK), Auxilin 2	Uncoating of clathrin coated vesicles	Risk locus	Risk factor candidate	n/a	Sporadic PD	[38,39]
DNAJC13	Receptor-mediated endocytosis 8 (RME8)	Retromer tubule formation, regulation of endosomal clathrin dynamics,	p.N855S	Autosomal dominant	67	Late onset PD, slow disease progression, good resonse to L-DOPA	[31,85–88]
DNAJC10	Endoplasmatic reticulum DNAJ domain-containing protein 5 (Erdj5)	Reductase involved with ER-associated degrataion of misfolded proteins	p.L301I	Risk factor candidate	n/a	Sporadic PD	[31]
DNAJC5	Cystein strint protein a (CSPa)	Synaptic protein folding	p.L115R, p.L116del	Autosomal dominant	25–46	Adult cerebroid neuronal lipofuscinosis, with generalized seizures, movement disorders, cognitive detoriation, progressive dementia; parkinsonisms obseved in some individuals	[33–35]
DNAJC12	DNAJC12	Interaction with aromatic amino acid hydrolases	p.K63*, p.V27Wfs*14	Autosomal recessive	13–51	Non-progressive, L-DOPA responsive parkinsonisms	[32]

HSPs were originally defined by their function in response to cellular stress, acting to regulate correct protein folding and chaperoning [4,5]. As more of the fundamental biology of the multiple HSP families has been uncovered, it is now clear that the HSPs encompass both stress-induced family members as well as constitutively expressed members playing critical roles in housekeeping functions [6]. HSPs are involved with the correct folding and refolding of polypeptides, thereby preventing them from forming dysfunctional proteins or potentially cytotoxic aggregates. When correct folding of the polypeptides is not possible, HSPs facilitate degradation either by preventing polypeptides from assembling or by facilitating their transport to degradative systems [7,8]. Since the accumulation of intracellular proteinaceous aggregates, chiefly composed of α-synuclein, is a pathological hallmark of PD, HSPs have emerged as a potential therapeutic target, as they have been shown to prevent and alleviated α-synuclein-mediated toxicity (reviewed in [9,10]).

Over a hundred known HSPs are classified into 8 different classes based upon domain organization and function [6,11]. The HSPA/HSP70 superfamily is a ubiquitous class of chaperones. In addition to their key role in protein assembly, HSP70 proteins have been implicated in a wide spectrum of cellular processes driven by protein interactions, including synaptic transmission, endoplasmatic reticulum (ER) stress response and endomembrane trafficking [6]. At the molecular level, HSP70 proteins have a remarkable sequence identity within and across species. They contain an N-terminal ATPase domain, a substrate binding domain and a C-terminal domain for nucleotide exchange-dependent substrate binding regulation. Under normal physiological conditions, HSP70 proteins constitute compartment-specific machineries, that inevitably also contain a J-protein for its function [12,13].

The J-protein (DNAJ/HSP40) family is, with over 40 members, the largest class of HSPs and is involved in the assembly and disassembly of macromolecular complexes [6,11]. DNAJ proteins function as co-chaperones and are responsible for much of the functional diversity of HSP70 proteins [14]. The DNAJ proteins interact with HSP70 through their highly conserved J-domain (Figure 1), and contain specific peptide epitopes for association with different substrates (Figure 2A) [13,15–17]. Understanding the molecular function of these proteins has been aided by progress in elucidating the atomic resolution structures of a number of members of this family, as shown in figure 2B. The expression of DNAJ proteins is often limited to specific tissues and cell types (as shown in Figure 1) or confined to subcellular compartments - therefore conferring specificity by recruiting HSP70 to defined protein complexes [14]. Since the intrinsic ATPase activity of HSP70 proteins is low, DNAJ proteins play a crucial role in stimulating Hsc70 to harness its ATPase activity required for substrate conformation [12,18–20]. The hydrolysis of ATP to ADP is required for the stabilization of the interaction, resulting in conformational changes of the substrate proteins (Figure 3) [13,19,20]. As ADP-bound HSP70 has a very slow engagement with client proteins, replacement of ATP for ADP by nucleotide exchange factors (NEFs) enables recycling of HSP70 molecules [12,21,22] (Figure 3).

Figure 1:

Overview of DNAJ protein phylogeny and expression. (A) Phylogenetic tree generated through neighbor joining, indicating cladistic relationship between DNAJC proteins. The gene tree was generated using the Simple phylogeny software [114], using default tree format, no distance correction and no exclusion of gaps. (B) Protein Blast of DNAJC proteins involved with parkinsonisms using the Constraint-based Multiple Alignment Tool [115], with indication of the conserved J-domain in red. (C) Overview of tissue expression of DNAJC proteins associated with PD based on the GTEx RNA sequencing database [56]. Expression levels are shown in TPM (transcripts per million kilobase) and counts are calculated from a gene model with isoforms collapsed into a single gene. Gene expression levels in different tissues are indicated by different colors, with yellow violin plots (boxed) indicating expression levels in brain tissue.

Figure 2:

Domain and protein structures of the DNAJC proteins. (A) Domain organization of DNAJC proteins with indication of pathogenic mutations. (B) Extant structural data for the DNAJC proteins. Clockwise from top left: i) The Auxilin (DNAJC6) in complex with clathrin, with inset showing the Auxilin J domain (blue) interacting with clathrin; ii) Protein structure of CSPα (DNAJC5); iii) Protein structure of ERdj5 (DNAJC10), with the J domain highlighted in green; iv) The dimeric structure of the GAK (DNAJC26) kinase domain. Images were derived from PDB co-ordinates 1xi5, 2n04, 4o38 and 3apo, respectively [116–119]).

Figure 3:

Model of DNAJ-assisted conformational protein changes and protein folding by HSP70. 1) DNAJ is recruited to and transiently interacts with client proteins (unfolded protein to be folded or folded protein to undergo conformational change). 2) HSP70 recruitment to the client protein is mediated by interaction of HSP70-ATP with the J-domain of DNAJ. 3) DNAJ stimulates the ATPase activity of HSP70-ATP to induce the conformational change of the client protein and DNAJ is released from the complex. 4) NEF protein is recruited to HSP70-ADP for nucleotide exchange of ADP to ATP. 5) HSP70-ATP is released to undergo a new round of protein folding.

The DNAJ protein family can be further subdivided into three classes (DNAJA, DNAJB and DNAJC) based on the presence and location of protein domains. Whereas DNAJC proteins are characterized by just the J-domain (Figure 1), DNAJA and DNAJB proteins contain an additional cysteine repeat domain and G/F domain or a G/F domain only, respectively [6]. Multiple members of the DNAJC family have been associated with PD and other neurodegenerative diseases displaying parkinsonisms (Table 1) [23,24,33–35,25–32]. This review will aim to explore the function of the DNAJC proteins and how the disease associated DNA variants in DNAJC proteins may contribute to dopaminergic degeneration and disease pathogenesis, recently reviewed in Gorenberg and Chandra, in the light of our increasing understanding of the human genetics underpinning this association [36].

2. Clathrin dynamics, DNAJ proteins and Parkinson’s disease

Vesicular transport is thought to play a prominent role in PD pathogenesis, as multiple genes involved with endomembrane trafficking underlie monogenic - often familial - PD (SNCA, LRRK2, VPS35 and ATP13A2) [37] and genome-wide association studies (GWAS) have nominated common risk variants contributing to sporadic PD (SNCA, LRRK2, GBA and RAB29) [38,39]. Remarkably, multiple genes involved with clathrin dynamics have been associated with PD, including synaptojanin 1 (encoded by SYNJ1) as well as several members of the DNAJC protein family (DNAJC6, DNAJC26 and DNAJC13) [23,24,26,38,40–42].

The cargo for clathrin-dependent vesicular transport is concentrated and selected by clathrin adaptor proteins [43]. Clathrin triskelia, consisting of three clathrin heavy chains and light chains, bind to the clathrin adaptor proteins to form clathrin coated pits [44]. Clathrin-coated pits can be formed from either the trans-Golgi network (TGN) or plasma membrane. As the vesicle matures, the clathrin triskelia assemble into the characteristic clathrin lattice to form a clathrin coated vesicle (CCV). Once the CCV is pinched off the membrane by dynamin, it requires uncoating of the clathrin coat in order to be able to fuse with its destination compartments. The disassembly of the clathrin coat is carried out by either Auxilin (encoded by DNAJC6) or GAK (encoded by DNAJC26) [45,46], whereas synaptojanin 1 mediates the shedding of accessory clathrin proteins from the vesicular membrane [47,48]. Finally, flat clathrin structures are present on endosomal subdomains. RME8 (encoded by DNAJC13) is thought to play a regulatory role in clathrin dynamics of those microdomains to allow for the formation of retromer tubules (Figure 4) [49].

Figure 4:

Overview of clathrin trafficking and clathrin dynamics in the cell, with indication of the involvement of DNAJ proteins. 1) Fission of clathrin coated vesicles from the trans-Golgi network or plasma membrane by dynamin 1, aided by DNAJC5. 2) Uncoating of clathrin coated vesicles by DNAJC6/Auxilin or DNAJC26/GAK. 3) Retromer-dependent recycling of cargo from endosomes mediated through DNAJC13/RME-8. 4) Exocytosis of vesicles at the plasma membrane, mediated through the SNARE protein SNAP25 and DNAJC5.

2.1. DNAJC6 (Auxilin) and recessive juvenile/early onset Parkinson’s disease

Auxilin, encoded by DNAJC6, is a neuronal co-chaperone of Hsc70/HSPA8 and plays a central role in the uncoating of CCVs [45]. At the protein level, Auxilin consists of an N-terminal PTEN-like domain for lipid binding, a domain with clathrin binding boxes and a C-terminal J-domain for Hsc70 binding (Figure 2). Auxilin is recruited to CCVs through interaction with the clathrin coat and phosphatidylinositol in the membrane of the vesicle, derived from either the plasma membrane or the Golgi apparatus [50–52]. The J-domain of Auxilin subsequently recruits Hsc70 to the CCVs and stimulates its ATPase activity to direct the uncoating of the clathrin coat (Figure 4) [53–55].

Two transcript isoforms of DNAJC6, with alternative splicing of exon 1 (ENST00000371069, ENST00000395325), are highly expressed in the brain (Figure 1) [56]. Multiple autosomal recessive loss of function mutations have been described to cause juvenile (age at onset <21 years) and early-onset (age at onset < 41 years) PD (Table 1) [23–26]. For simplicity, the mutations referred to hereafter are described in relation to ENST00000371069 (NM_001256864), as this is considered to be the canonical isoform (Figure 2).

The homozygous splice acceptor site DNAJC6 variant c.801–2A>G was first described in 2012 to cause juvenile onset PD in a consanguineous family of Palestinian origin [23]. The mutation lies within the exon-intron boundary at the start of exon 7 and has been shown to result in mis-spliced mRNA and overall decreased mRNA levels [23]. A homozygous synonymous p.T741= variant was discovered in a Brazilian family with early onset PD. This variant is located 5 bases before the end of exon 15 and results in mis-splicing with subsequent decreased mRNA levels [25]. In addition, two C-terminally truncating mutations that completely lack the J-domain, p.Q791* and p.Q846*, were described in consanguineous families of Turkish and Sudanese/Yemini origin with juvenile PD, respectively [24,26]. Finally, p.R927G, a homozygous missense mutation in the highly conserved J-domain of Auxilin, originates from a Dutch proband and was found to cause early onset PD [25].

Although there is a relatively small number of Auxilin variant carriers, detailed clinical evaluation supports a genotype-phenotype correlation. Patients with missense mutations in DNAJC6, such the R927G variant, present with early-onset parkinsonism. Patients with splicing mutations (c.801–2A>G, p.T741=) presented both juvenile and early onset PD [23,25]. In contrast, nonsense mutations (p.Q791*, p.Q846*) result in a very severe and rapidly progressing disease course with juvenile onset [24,26]. These patients also presented additional atypical features including cognitive impairment, pyramidal signs and, in some cases, hallucinations and seizures. Remarkably, a homozygous 80kb deletion encompassing the C-terminus of DNAJC6 as well as LEPR has previously been associated with epilepsy and mental retardation in the absence of any parkinsonian phenotype [57]. However, the patient was 7 years old and clinical presentation possibly incomplete. Response to L-DOPA in patients with DNAJC6 mutations is either absent or limited due to L-dopa induced hallucinations and dyskinesias [23–26].

Sporadic patients with early onset PD have been identified with compound heterozygous mutations (c.203813A>G and c.1468183del) or heterozygous variants (p.L209P, p.R619C, p.M133L, p.F839Lfs*22) in DNAJC6 [25]. However, this observation has not been replicated in larger studies and so it currently unclear as to whether heterozygous variants in DNAJC6 are causal or a risk factor for disease. Whereas the role of Auxilin in the uncoating of CCVs is well known, the mechanism of action of the protein in disease pathogenesis remains to be elucidated. The underlying genetics of DNAJC6 point to a loss of function mechanism for Auxilin in the etiology of parkinsonism, possibly resulting in impairment of uncoating of CCVs at the Golgi or the synapse. Indeed, depletion of Auxilin results in an accumulation of clathrin structures and impaired delivery of cargo [58–63]. Complete loss of Auxilin at the organismal level has been found to result in neurodegeneration and locomotor deficits in drosophila and synaptic defects in mice [61,64]. In addition, expression of Auxilin was found to partially restore synaptic function in patient-derived iPSC neurons [65].

2.2. DNAJC26 (GAK) as risk factor for Parkinson’s disease

Whereas DNAJC6 expression is mainly restricted to brain, DNAJC26, encoding GAK (cyclin G-dependent kinase A), is ubiquitously expressed with lower relative expression levels in brain (Figure 1) [56]. GAK shares 44% identity and 59% homology to Auxilin and has a similar domain structure, with an additional N-terminal kinase domain (Figure 2), responsible for most of the variability at the sequence level [66]. Like Auxilin, GAK functions as a co-chaperone of Hsc70 in the uncoating of CCVs (Figure 4) [46]. Knockout (KO) of DNAJC26 in the mouse is embryonic lethal, whereas DNAJC6 KO animals have an increased postnatal mortality and decreased birth weight but otherwise normal lifespan [67]. The GAK locus has been identified by GWAS to be a risk factor candidate for PD [38,39]. However, this region contains several other genes of interest, including TMEM175, a lysosomal gene that has been demonstrated to interact with α-synuclein at a functional level, therefore it is currently unclear which of these genes is the causal gene in this locus [27,68,69]. This region on chromosome 4 is further complicated by the presence of two independent risk signals for PD, and due to GAK and TMEM175 sharing the same promoter [39]. Future research is required to clarify how this locus contributes to the pathogenesis of PD.

2.3. DNAJC13 (RME-8) and Parkinson’s disease

RME-8 (receptor mediated endocytosis 8) was initially identified through a genetic screen for endocytic deficits in the nematode model organism C. elegans [70]. RME-8 consists of a J-domain flanked by two IWN repeat domains at either side (Figure 2) [71]. Similar to Auxilin and GAK, RME-8 confers specificity to the Hsc70 chaperone through its J-domain [71–74]. Rather than uncoating CCVs, RME-8 has been found to regulate clathrin dynamics on early endosomes [75,76] (Figure 4).

Endosomes are sorting hubs wherein proteins are sorted to the lysosomes for degradation, trafficked to the plasma membrane, or recycled to the TGN via the retromer. The retromer is a multimeric protein complex that recycles proteins for retrograde transport by concentrating cargo in tubules formed from the endosome [77,78]. The formation of tubules is controlled by the SNX dimer component of the retromer, whereas cargo selection is mediated by a trimer of VPS26, VPS29 and VPS35 [78,79]. Notably, multiple coding mutations in VPS35 have been demonstrated to cause monogenic PD [80,81]. In addition to those core components, a number of accessory proteins are recruited to the retromer, including the WASH complex, enabling the formation of a branched actin network to enable cargo sorting. RME-8 has been found to interact both with SNX1, a component of the SNX dimer, as well as FAM21, a protein from the WASH complex [75,82].

RME-8 is recruited to early endosomes through interaction with PI(3)P [72,83]. RME-8 does not appear to bind to clathrin directly, but loss of either RME-8 or Hsc70 results in an accumulation of clathrin at early endosomes [71,75,84]. In addition, loss of RME-8 results in the missorting of retromer-dependent cargo [70,74–76,84].

Mutations in DNAJC13 were originally described in a large multi-incident family with Dutch-German-Russian Mennonite ancestry presenting with autosomal-dominant inheritance of PD [85]. A p.N855S coding mutation was identified in multiple cases in this family. The association of the p.N855S DNAJC13 mutation with PD is further supported by the absence of this mutation in controls, identification of 5 additional PD cases carrying p.N855S mutation and high level of conservation through evolution of the mutated amino acid. However, the p.N855S mutation does not appear to be a common cause for PD, as other data from Caucasian and Chinese cohorts with PD in four additional independent studies only identified one case with p.N855S mutation, sharing the same haplotype as the original Mennonite family [31,86–88]. In addition, a number of DNAJC13 variants of unknown pathogenicity have been suggested to contribute to susceptibility of disease development, including p.P336A, p.V722L, p.R1266Q, and p.T1895M [89]. Affected DNAJC13 p.N855S carriers present typical late onset PD with mean age of onset of 67. Onset of symptoms was asymmetric and disease progression was slow, with a combination of cardinal PD features (tremor, rigidity and bradykinesia) and a good response to levodopa [42]. Interestingly, the p.N855S mutation has also been observed in patients with essential tremor [90].

However, it is important to note that there have been conflicting reports on association of p.N855S mutation with PD. An recent independent re-analysis of the original family has pointed to a variant in TMEM230 to be causative for disease in this family [91]. Importantly, the suggested variants in both DNAJC13 and TMEM230 have imperfect disease segregation, thus further investigations are necessary to shed light on this complex genetic situation.

3. DNAJC10 (ERdj5) and Parkinson’s disease

Erdj5, encoded by DNAJC10, is a ubiquitously expressed (Figure 1), endoplasmic reticulum (ER)-resident oxidoreductase [92]. In eukaryotic cells, proteins are co-translationally transported through the ER and acquire posttranslational modifications such as disulfide bonds between cysteine residues for the stabilization of the tertiary protein structure. As an ER quality control mechanism, proteins that are terminally misfolded are retro-translocated from the ER to the cytosol through the ER-associated degradation (ERAD) pathway for degradation by the ubiquitin-proteasome system. DNAJC10 is recruited to misfolded proteins targeted for degradation through interaction with proteins of the ERAD machinery [93–95]. DNAJC10 then executes the reduction of the disulfide bonds, required for subsequent passage of the misfolded proteins through the ER membrane. In addition to its role in ERAD, DNAJC10 is also required for the correct folding of substrates in the ER through the reduction of non-native disulfide bonds [96]. DNAJC10 interacts through its N-terminal J-domain with the ER-resident Hsp70 homologue HSPA5 (BiP) [92,93]. The role of DNAJC10 in degradation and folding in the ER depends on the interaction with the ER-resident Hsp70 homologue BiP through its N-terminal J-domain, as well as its disulfide exchange reaction through the thioredoxin domains [93,96].

A missense variant (p.L301I) in DNAJC10 was found to be associated with a decreased risk of PD in a cohort of 512 patients and controls in the Chinese Han population [31]. However, the protective association of this variant with PD is to be interpreted with caution as it has not been replicated in an independent cohort, and no association in this region was identified in the most recent large-scale GWAS [38,39].

4. DNAJC12 and non-progressive parkinsonisms

DNAJC12 encodes a small protein with high relative expression levels in the brain (Figure 1). At the protein level, DNAJC12 contains the characteristic J-domain at the N-terminus, but has no other known functional domains (Figure 2). Mutations in DNAJC12 have recently been associated with recessive, early-onset, non-progressive parkinsonism in two kindreds of Canadian and Italian descent with an age of onset between 13 and 53 years old [35]. Exome sequencing identified homozygous loss of function variants, the nonsense mutation p.K63* in the Canadian family and the splicing mutation p.V27Wfs*14 in the Italian family [32]. Both mutations were absent in 497 patients with early-onset PD and 702 sporadic PD patients in the Chinese Han population, suggesting that mutations in DNAJC12 are not a common cause for PD [32,97]. Mutation carriers show mild, non-progressive parkinsonism and cognitive impairment, and are responsive to levodopa. Mutations in DNAJC12 were previously described in children with hyperphenylalaninemia, intellectual disability and dystonia [98]. Although the exact function of DNAJC12 is unknown, it is thought to interact with aromatic amino acid hydrolases, including tyrosine hydroxylase and tryptophan hydroxylase, involved in the synthesis of the neurotransmitters dopamine and serotonin, respectively. Both mutations result in severely decreased mRNA levels of DNAJC12 and are therefore thought to result in at least partial loss of function. Neuropathological analysis of the Canadian proband showed no signs of neurodegeneration or α-synuclein pathology and DNAJC12 mutations could confer secondary parkinsonism through a distinct etiology. As the substantia nigra appeared to be hypopigmented, disease mechanism could be explained by a chronic deficit of dopamine, consistent with the non-progressive course of the disease [32].

5. DNAJC5 and neuronal ceroid lipofuscinosis with parkinsonism

Neuronal ceroid lipofuscinoses (NCL) are a heterogeneous group of hereditary, progressive neurodegenerative diseases, characterized by an aberrant accumulation of auto-fluorescent pigment in the lysosomal compartment. Whereas NCL mainly affects children, mutations in a handful of genes, including DNAJC5, have been reported to cause adult onset NCL [35,99].

DNAJC5 encodes cysteine string protein α (CSPα). CSPα is abundantly present in presynaptic vesicles to assure correct folding of synaptic proteins through interaction with Hsc70 [100–103]. The SNAP (Soluble NSF Attachment Protein) Receptor, or SNARE, protein SNAP25, involved with the fusion of synaptic vesicles, is degraded by the proteasomal system unless assisted by CSPα for correct folding [103,104]. Remarkably, PD associated protein α-synuclein was found to assist CSPα in correct SNARE complex assembly [105]. CSPα has also been found to interact with dynamin, an important protein in the fission of CCVs [101].

In addition to its N-terminal J-domain, CSPα contains a cysteine string domain required for membrane binding and undergoes palmitoylation required for the correct palmitoylation-dependent sorting of CSPα. Two neighboring leucine residues in the CSPα cysteine string domain are hotspots for mutations in NCL, as independent groups have characterized the point mutation p.L115R and in-frame deletion p.L116del in multiple pedigrees to cause autosomal dominant adult-onset NCL. In addition, both mutations were identified in seemingly sporadic cases as well [33,34]. Affected patients have generalized seizures, dyskinesias, progressive cognitive deterioration and psychiatric manifestations [33–35]. In addition, some of the p.L166del carriers of a pedigree from Alabama were reported as presenting with parkinsonism [34,106]. Both DNAJC5 mutations were found to impair correct palmitoylation and membrane binding, resulting in mis-localization of CSPα. Overall, levels of CSPα were found to be decreased in neurons of affected individuals [33]. CSPα appears to be biochemically linked to another NCL-associated gene PPT1 (palmitoyl-protein thioesterase 1), as PPT1 is upregulated, mislocalized and hypo-active in the brains of CSPα mutation carriers, altogether suggesting a functional link between NCL-associated genes [107].

6. Conclusions

In the two decades since the identification of the first monogenic form of PD, caused by mutations in SNCA, substantial progress has been made in understanding the genetics of this disorder [108]. Approximately 1–5% of all PD cases are explained by monogenic mutations, and over ninety risk factors have been nominated by GWAS to contribute to sporadic PD [38,39]. It has been proposed that there are functional relationships between a number of these genes, and that multiple genetic risk factors may converge on a handful of pathobiological pathways. Importantly, as idiopathic PD cases share clinical and pathological features with familial forms of PD, the pathobiology of the more common sporadic form of disease may be explained by similar underlying mechanisms [109]. In the broader context of movement disorders, although PD is the most common clinical manifestation of parkinsonism, accounting for 85% of all cases, there are a range of additional parkinsonian syndromes that further expand the clinical and genetic spectrum of these disorders and have provided important mechanistic clues as to the changes that lead to neurodegeneration [2].

Over the past ten years, multiple members of the DNAJC protein family have been associated with PD and other parkinsonian disorders. The DNAJC family is a diverse class of proteins important in a wide spectrum of cellular functions. Remarkably, the DNAJC genes with strong genetic evidence for a link to parkinsonisms (DNAJC6, DNAJC5 and DNAJC12) display relatively high and specific expression in brain compared to other tissues (Figure 2), underscoring their important roles in healthy brain and neuronal function.

It should be noted that some of the reported genetic associations of DNAJC genes (DNAJC26, DNAJC13, DNAJC10) with PD and parkinsonism require additional genetic evidence to link the mutations with disease pathogenesis. First, the PD risk locus containing DNAJC26 (GAK) contains two independent variants in GAK as well as TMEM175 [38,39]. The current evidence is inconclusive as to whether GAK or TMEM175 is the causative gene at this locus [39]. More research is needed for a definite association of either or both genes with PD. Secondly, conflicting results have been reported in the pedigree in which DNAJC13 mutations were originally described to cause PD [86]. It remains uncertain whether mutations in DNAJC13 (located on chromosome 3) or TMEM230 (located on chromosome 20) are causative for disease, as mutations in neither of those genes show complete penetrance with disease [86,91]. More functional and genetic data will help to clarify this complex genetic situation. Finally, mutations in DNAJC10 that have been associated with PD have been described in a relatively small Chinese cohort [31]. These results will have to be replicated on a larger scale to definitely link DNAJC10 mutations to PD.

Multiple DNAJC proteins that are associated with PD (Auxilin/DNAJC6, GAK/DNAJC26, RME-8/DNAJC13) play important roles in the regulation of clathrin trafficking and dynamics. The association of another key gene in clathrin trafficking, SYNJ1, with PD and the finding that multiple PD genes interact with clathrin, further supports the importance of efficient clathrin dynamics for healthy neuronal function [29,40,41]. Mutations in the synaptic protein CSPα (DNAJC5) cause NCL displaying parkinsonism. In addition to its role in exocytosis, CSPα also converges on the clathrin trafficking pathway through interaction with dynamin 1 [101]. Remarkably, ATP13A2 is another gene that has been associated both with NCL and Kufor Rakeb Syndrome, a rare atypical parkinsonism, highlighting the interconnectivity of pathobiological pathways of the spectrum of parkinsonian syndromes [110,111]. Finally, DNAJC12 mutations have been associated with non-progressive parkinsonisms [32]. As this is likely a secondary parkinsonism, mutations in this gene may function through distinct pathobiological mechanisms.

Reviewing the functions of the DNAJC proteins associated with parkinsonisms and taking into account genetic evidence and gene expression levels in brain, highlights an important role for synaptic trafficking and clathrin dynamics in the pathogenesis of parkinsonisms and PD. How exactly these integrate or overlap with previously identified pathways leading to nigral cell death, however, is unclear. The potential for alterations in endocytosis of aggregated proteins, and in particular α synuclein, to impact on the hypothesized role of prion like spread of aggregates in PD is notable as one possible route, as is the possibility that changes in synaptic function may lead to loss of synapses, axonal degeneration and eventually neuronal cell death [112,113]. It is clear that further research is required in order to elucidate how mutations in DNAJC proteins and the subsequent misregulation of clathrin dynamics contributes to disease pathogenesis. It is likewise clear that developing a greater understanding of the molecular genetics of the DNAJC genes, and further dissecting the mechanisms underlying the full spectrum of parkinsonian disorders, will be critical for the development of pathway-targeted therapeutics to modulate disease progression in PD.

Abbreviations

PD

Parkinson’s disease

HSP

heat shock protein

ER

endoplasmic reticulum

NEF

nucleotide exchange factor

TGN

trans golgi network

CCV

clathrin coated vesicle

ERAD

endoplasmic reticulum associated degradation

GWAS

genome wide association study

NCL

neuronal Ceroid lipofuscinosis