Neddylation-dependent RAB40-Cullin5 E3 ligase complex controls ubiquitin-mediated proteasomal degradation of DNAJC13

Highlights

• •RAB40A/B/C bind to DNAJC13 via its N- and C-terminal IWN1 and IWN3 domains.
• •RAB40–CRL5 complex enhances DNAJC13 polyubiquitylation and promotes its proteasomal degradation.
• •Knockdown of RAB40A/B/C increases DNAJC13 levels and enhances autophagic flux.
• •Neddylation regulates RAB40-CRL5-mediated DNAJC13 degradation.

Abstract

DNAJC13 is a member of DNAJ co-chaperone family and its missense mutation causes a rare familial form of Parkinson’s disease. DNAJC13 is indispensable for neuronal homeostasis by regulation of autophagy and the regulation of receptor expression on the cell surface. However, the regulatory mechanisms governing DNAJC13 expression remain poorly understood. We previously demonstrated that DNAJC13 binds to RAB40 small GTPase. Unlike other RABs, RAB40 proteins interact with activated Cullin-5 (CUL5) via the SOCS box, forming an Cullin-RING E3 ubiquitin ligase (CRL) complex that regulates ubiquitylation of target proteins. Thus, as a binding partner of RAB40s, DNAJC13 degradation may be regulated by this functional complex. Here, we found that both N-and C-terminal regions of DNAJC13, including IWN1 and IWN3 domains, were essential for its binding to RAB40 A/B/C. Moreover, DNAJC13 degradation of relies predominantly on the ubiquitin-proteasome system rather than on autophagy, and upregulated DNAJC13 ubiquitination in the presence of RAB40A/B/C. Intriguingly, RNAi-mediated silencing of the RAB40 subfamily caused a significant increase in DNAJC13 expression, accompanied by higher autophagic flux. Furthermore, co-immunoprecipitation experiments demonstrated that RAB40C, which is predominantly expressed in midbrain dopaminergic neurons, binds to CUL5 and its activator NEDD8 as well as RBX2 (E2). Finally, DNAJC13 cellular levels were regulated by NEDD8 activity. Taken together, these findings indicate that NEDD8-mediated neddylation plays a key role in RAB40-CUL5-mediated DNAJC13 degradation.

Graphical abstract

1. Introduction

Parkinson's disease (PD) is a common movement disorder affecting 1% of individuals >60 years old [1]. The pathological hallmarks of PD are loss of dopaminergic neurons (DANeu) in the midbrain and the appearance of intra-neuronal inclusions called Lewy bodies, comprising mainly phosphorylated, misfolded α-synuclein (αS) [2]. Despite the possible involvement of multiple pathogenic cascades in PD, the endo-lysosomal system is emerging as a key player in its pathogenesis [3].

DNAJC13 (DnaJ homolog C, member 13), also known as PARK21, has been identified as the causative gene for the dominant form of familial PD with LB pathology [4,5]. DNAJC13, formerly designated as receptor-mediated endocytosis-8 (RME-8), was originally identified in a screen for endocytic defect phenotype in nematodes [6]. DNAJC13, a large protein with a molecular mass of 254 kDa, is ubiquitously expressed and abundant in the nervous system [7]. Structurally, it contains four IWN (isoleucine, tryptophan, and asparagine) repeats and central DnaJ domain (J-domain). DNAJC13 mainly localizes on the early endosome and interacts with the tail domain of FAM21 (family with sequence similarity 21) in the Wiscott-Aldrich and Scar Homolog (WASH) complex and sorting nexin dimer, key determinants for endosomal subdomain organization of retrograde trafficking (i.e., retromer) [[8], [9], [10], [11]]. As such, DNAJC13 plays a crucial role in endosomal trafficking by regulating endosomal recycling and autophagy-lysosomal degradation[[12], [13], [14], [15], [16]]. Previously, we identified DNAJC13 as a binding partner of the RAB40 family [17]. RABs function as molecular switch changing between active and inactive states [18]. Upon cellular demand, activated RABs localize to specific membrane domain and recruit distinct binding partners, thereby regulating specific membrane trafficking pathways. RAB40 originated in bilaterians as a result of RAB18 duplication [19]. After successive duplication events, RAB40 evolved into a subfamily of four proteins (i.e., RAB40A, RAB40B, RAB40C, and RAB40AL) in primates [20]. In contrast to other RABs, RAB40 proteins uniquely contain a SOCS box domain in the C-terminus through which they interact with the Cullin-RING E3 ubiquitin (Ub) ligase (CRL) complex, ubiquitinating their client proteins [20,21]. In this multisubunit complex, CRL5 serves as scaffold that recruits specific client proteins via RAB40 targeting them for degradation through the ubiquitin-proteasome system (UPS) [22,23].

Accordingly, we hypothesized that DNAJC13 is directed to CRL5-mediated degradation supported by RAB40s. In addition, we investigated whether NEDD8 (neuronal precursor cell-expressed developmentally downregulated protein 8)-mediated neddylation, a CRL5 activation process, functions upstream of this pathway [24].

2. Materials and methods

2.1. Plasmid preparation and transfection

EGFP-tagged human DNAJC13 expression vectors including wild-type (WT), ΔC425 mutant (1–5454 bp), ΔC633 mutant (1–4830 bp), ΔN453 mutant (1360–6729 bp), ΔN1083 mutant (3250–6729 bp), and N855S mutant were previously described [10,12]. The CUL5 expression vector (pcDNA3-myc-CUL5) was purchased from Addgene (Watertown, MA). mCherry-tagged human WT RAB40A, RAB40B, and RAB40C expression constructs (RAB40A-WT/pEGFP-C1, RAB40B-WT/pEGFP-C1, RAB40C-WT/pEGFP-C) [25,26], and Ub expression vector (HA-Ub/pEF) were previously described [27]. Site-directed mutagenesis was performed on each mCherry-RAB40-WT expression vector to substitute the four amino acid sequence of the Cul-box (LPLP) by alanine residues (AAAA). The UBE2F expression vector (FLAG-UBE2F/pIRES2) was kindly provided by Dr. Yi Sun (Zhejiang University School of Medicine, Hangzhou, China). Plasmids were transfected into human embryonic kidney 293 (HEK293) cells using the NEPA21® square wave electroporator (Nepa Gene, Chiba, Japan). For the autophagy flux assay, GFP-RFP-LC3B (P36239, Thermo Fisher Scientific) was used according to the manufacturer’s instructions.

2.2. Cell culture

HEK293 cells and SH-SY5Y cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) with high glucose (4500 mg/L) or DMEM with low glucose (1000 mg/L) (FUJIFILM Wako Chemicals, Osaka, Japan), supplemented with 4 mM L-glutamine and 10% fetal bovine serum (FBS; S1810–500, Biowest) at 37 °C incubator in humidified 5% CO₂/air. For the generation of stable cell lines, cells were cultured under the selection pressure of G418 sulfate (800 μg/mL; InvivoGen, San Diego, CA). Cycloheximide (CHX; Sigma, St. Louis, MO), MG132 (Cayman Chemical, Ann Arbor, MI), Bafilomycin A1 (Baf A1; Sigma), rotenone (Sigma), and MLN4924 (CST, Danvers, MA) were added to the culture medium at the concentrations indicated.

2.3. siRNA transfection

siRNAs targeting RAB40A, RAB40B, RAB40C were previously reported [26]. siRNA-A (sc-37,007, Santa Cruz, Heidelberg, Germany) was used as control. All siRNAs were introduced into cells using the NEPA21® electroporator.

2.4. Co-immunoprecipitation

Cells were lysed in lysis buffer ( 50 mM HEPES-KOH [pH 7.2], 150 mM sodium chloride, 1 mM magnesium chloride, 0.5% Triton X, and 1×cømplete® protease inhibitor cocktail; Roche, Basel, Switzerland). Then, lysates were incubated overnight at 4 °C with anti-GFP mAb-magnetic beads (D153–11, MBL, Tokyo, Japan), anti-RFP mAb-magnetic beads (M165–11, MBL), or anti-myc-tag mAb-magnetic beads (M047–11, MBL). After washing three times with lysis buffer, the proteins were eluted using 2x Laemmli buffer at room temperature for 90 min before Western blotting (WB).

2.5. Western blot analysis

Cell lysates denatured with Laemmli buffer were separated by SDS-PAGE and transferred onto a polyvinylidene difluoride membrane (Immobilon-P; Merck Millipore). Then, the membrane was blocked with either 5% skim milk dissolved in tris-buffered saline with 0.1% Tween 20 (TBST) or Can Get Signal immunoreaction enhancer solution (NYPBR, Toyobo, Osaka) and incubated overnight at 4 °C with the following primary Abs: anti-β-tubulin (66,240-1-Ig, 1:20,000; Proteintech, Rosemont, IL), anti-GFP (598, 1:20,000; MBL), rabbit anti-mCherry (226,765, 1:1000; Proteintech), goat anti-mCherry (NBP3–05,558, 1:2000; NOVUS, Centennial, CO), anti-GAPDH (14C10, 1:2000; CST), anti-myc (SM1535P, 1:1000; Acris), anti-HA (#2367, 1:1000; CST), Anti-RAB40C (PA5–104,080, 1:2000; Invitrogen), Anti-FLAG (F1804, 1:1000; Sigma), Anti-RNF7 (RBX2) (EPR12001, AB181986, 1:1000; Abcam, Cambridge, UK), Anti-NEDD8 (34–1400, 1:1000; Invitrogen). The human anti-DNAJC13 antibody was already described [10] and obtained from Eurofins Genomics (targeting 1965–2243aa). After washing the membrane three times with TBST, the membrane was incubated with HRP-conjugated secondary Abs for 1 h under gentle shaking. Next, bands were visualized with Luminata Forte Western HRP substrate (Millipore) and images were captured by the ChemiDoc™ MP imaging system (Bio-Rad, Darmstadt, Germany). Band intensities were quantified using ImageJ software (NIH) or the ChemiDoc™ MP imaging system (Bio-Rad).

2.6. Fluorescence confocal microscopy

Fluorescent images of unfixed cells expressing GFP- and RFP-tagged proteins were acquired and analyzed using a confocal laser scanning microscope system (Fluoview FV300, OLYMPUS).

2.7. Expression profile analysis using postmortem brain and bulk tissue dataset

The postmortem brain dataset was obtained from Single Cell PORTAL (https://singlecell.broadinstitute.org/single_cell/), and cell annotations were based on the identifiers registered on [28]. After normalizing the dataset, principal component analysis was performed on the matrix, and the top 30 principal components were embedded onto two dimensions using the Uniform Manifold Approximation and Projection (UMAP) algorithm with the number of neighbors set to 30, following the default implementation of Seurat v5. The RNA expression levels of the RAB40 subfamily across different nervous tissues were obtained from the publicly available dataset of the Human Protein Atlas (https://www.proteinatlas.org/). Expression levels are estimated as normalized transcripts per million (nTPM), calculated by integrating RNA-seq data from the GTEx Consortium (https://gtexportal.org/home/) and Human Protein Atlas (https://www.proteinatlas.org) datasets, with normalization performed across tissues. These values represent bulk tissue–level gene expression.

2.8. Statistical analysis

Statistical analyses are performed using R software (R Foundation for Statistical Computing, Vienna, Austria). Graphs are visualized using ggplot2 (https://ggplot2.tidyverse.org), and the quantitative data are expressed as the mean ± SE.

3. Results

3.1. DNAJC13 interacts with RAB40 proteins

Previously, we demonstrated that human RAB40A, mouse Rab40B, and mouse Rab40C interact with mouse DNAJC13 [17]. RAB40AL exhibits a high similarity to RAB40A, with 98% identity at amino acid level (UniProt ID: P0C0E4, Q8WXH6); however, its transcript was hardly detectable in DANeu and therefore excluded from subsequent experiments. Using human bait and prey proteins, DNAJC13 co-immunoprecipitated with human RAB40A/B/C and vice versa (Fig. 1A and B and Suppl. Fig. 1A and 1B). This result was confirmed by confocal microscopy (Fig. 1C). Since the IWN repeats in DNAJC13 play important roles in interactions with client proteins such as SNX1 [29], we investigated whether mutant DNAJC13s lacking IWN domain(s) (Fig. 1D) lost binding to RAB40s. Strikingly, among four deletion mutants, ΔC633and ΔN1083 failed to interact with RAB40A/B/C (Fig. 1E–G and Suppl. Fig. 1E - G), suggesting that both the N-and C-terminal segments of DNAJC13, including IWN1 and IWN3, are essential for binding to RAB40s. We also examined the effect of the PD-associated pathogenic N855S mutation in DNAJC13 on its interaction with RAB40C by co-immunoprecipitation; however, no appreciable reduction or loss of interaction was observed (Suppl. Fig. 8).

Fig. 1.

DNAJC13 interacts with the RAB40 subfamily. (A, B) Co-immunoprecipitation assays in HEK293 cells show that human DNAJC13 physically associates with human RAB40A/B/C and vice versa, with no substantial differences in binding affinity among the RAB40 isoforms. (C) Confocal microscopy demonstrating subcellular colocalization of EGFP-DNAJC13 and mCherry-tagged RAB40 proteins in HEK293 cells. Scale bar: 20 μm. (D) Schematic representation of EGFP-tagged wild-type (WT) and four deletion mutants of human DNAJC13. Black boxes indicate individual IWN domains, and the red box denotes the conserved J-domain. (E–G) Among the four mutants, ΔC633 and ΔN1083 fails to interact with RAB40A/B/C, indicating that both the N- and C-terminal segments of DNAJC13, including the IWN1 and IWN3 domains, are required for RAB40 binding.

3.2. RAB40 interacts with CUL5 through its cul5 box

RAB40s serve as adaptor protein for CUL5, a CRL5 component, regulating ubiquitylation of specific target proteins. Structurally, the SOCS box consists of BC and Cul5 boxes, which interact with Elongin BC and CUL5, respectively (Fig. 2A). Particularly, the LPLP sequence in the Cul5 box is indispensable for binding with CUL5 [30]. To confirm that the LPLP motif mediates binding to CUL5 in our experimental model, we performed a co-immunoprecipitation assay using a RAB40 mutant in which the LPLP motif within the SOCS box was replaced with alanine residues (LPLP to AAAA), designated RAB40^4A. As expected, the amino acid substitution in RAB40A/B/C abrogated interaction with CUL5 (Fig. 2A and B and Suppl. Fig. 2B), while the interaction with DNAJC13 remained unchanged (Fig. 2C and Suppl. Fig. 2C) as did the expression level of EGFP-DNAJC13 in RAB40^WT and RAB40^4A expressing cells.

Fig. 2.

The RAB40 subfamily interacts with CUL5 through the Cul5 box. (A) Schematic showing the C-terminal region of human RAB40C, containing a SOCS box with BC and Cul5 boxes (dotted lines) harboring the LPLP motif (highlighted in yellow). (B) Mutation of the LPLP motif to AAAA disrupts the interaction between RAB40 proteins and CUL5. (C) This mutation does not affect the RAB40-DNAJC13 interaction. EGFP-DNAJC13 expression levels are comparable between RAB40^WT- and RAB40^4A-expressing cells.

3.3. RAB40C is highly expressed in DANeu

To delineate the expression profile of RAB40 subfamily in the human midbrain, a single-cell RNA sequencing (scRNA-seq) database in postmortem brain was used. Cells were plotted according to the coordinates obtained by dimension reduction using UMAP (Fig. 3A). Cell identification was based on the annotations provided by the database, rather than on manual cluster designation. Expression of RAB40s in cells was evaluated by UMAP or violin plots. RAB40B expression was abundant in oligodendrocytes (Olig) (Fig. 3B and C). In contrast, RAB40C was predominantly expressed in DANeu, and to a lesser degree, in non-dopaminergic neurons (Neu). Both RAB40A and RAB40AL were expressed at low levels in human midbrain tissue. In addition, we evaluated the RNA expression levels (nTPM) of the RAB40 subfamily across human nervous tissues using publicly available dataset of the Human Protein Atlas (https://www.proteinatlas.org/). In the central nervous system, RAB40 family exhibits distinct expression patterns, i.e., RAB40A shows relatively high expression in the basal ganglia, RAB40AL in the cerebellum, RAB40B in the basal ganglia and spinal cord, and RAB40C in the cerebral cortex, basal ganglia, and cerebellum (Suppl. Fig. 9).

Fig. 3.

Expression profiles of RAB40 subfamily in postmortem midbrain tissues.In silico analysis using publicly available transcriptomic data from aged healthy midbrain samples. (A) Cell clustering visualized by two-dimensional dimensionality reduction. Cell-type identities are shown in the right panel. (B) Expression patterns of individual RAB40 genes mapped onto the cell projection. (C) Violin plots showing expression intensities and cell ratios of each RAB40 family member across cell clusters. Abbreviations: Neu, neurons; DANeu, dopaminergic neurons; Astro, astrocytes; Oligo, oligodendrocytes; OPC, oligodendrocyte precursor cells; Mgl, microglia; Endo, endothelial cells.

3.4. DNAJC13 is degraded through the ubiquitin-proteasome system

As a binding partner of RAB40s, DNAJC13 was hypothesized to undergo ubiquitination through the CRL5 complex. To test this possibility, WT EGFP-DNAJC13, HA-Ub, and mCherry-RAB40s were co-expressed in HEK293 cells and the ubiquitination level of immunoprecipitated DNAJC13 examined by WB analysis. Remarkably, anti-HA immunoreactivity ladders (indicative of polyubiquitination) were observed in EGFP-DNAJC13 immunoprecipitates; this effect was potentiated by co-expression of RAB40s. Note that HA-Ub signals were observed under RAB40 expression (Fig. 4A and Suppl. Fig. 4A). However, no significant difference in ubiquitination were observed among members of the RAB40 subfamily (Fig. 4B). Next, a CHX chase assay was performed to determine whether RAB40C-induced ubiquitination promotes DNAJC13 degradation. After prolonged incubation (12 h), DNAJC13 degradation significantly increased in the presence of RAB40C (Fig. 4C and D and Suppl. Fig. 4C). Finally, we verified the impact of chemical interference with the UPS or autophagy on DNAJC13 degradation. MG132 (an UPS inhibitor) but not BafA1 (an autophagy inhibitor) significantly increased DNAJC13 expression level over 12 h (Fig. 4E–H, and Suppl. Fig. 4E and G). Likewise, MG132 stabilized DNAJC13 expression even with translation inhibition with CHX (Fig. 4I and J, and Suppl. Fig. 4I).

Fig. 4.

DNAJC13 degradation through the ubiquitin-proteasome system is facilitated by RAB40. (A, B) HEK293 cells expressing EGFP-DNAJC13, HA-Ub, and mCherry-RAB40 proteins are subjected to immunoprecipitation and Western blotting. High-molecular-weight HA-positive ubiquitin ladders indicate polyubiquitylation, which is enhanced by RAB40 co-expression without significant isoform-specific differences. Graphs show mean ± SEM. *p < 0.05 (one-way ANOVA with Tukey’s post-hoc test; n = 3). (C, D) Cycloheximide (CHX; 50 μg/mL) chase assay. Cells over-expressing RAB40C accelerates DNAJC13 decay over 12 h compared to Mock (mcherry) transfectant. (E–H) MG132 (1.25 μM; proteasome inhibitor), but not Bafilomycin A1 (BafA1, 100 nM; autophagy inhibitor), increases DNAJC13 protein levels over 12 h. (I, J) MG132 stabilizes DNAJC13 even during CHX-induced translation blockade. Graphs show mean ± SEM. *p < 0.05, **p < 0.01 (one-way ANOVA with Tukey’s post-hoc test).

3.5. RNAi-mediated silencing of RAB40 slightly increases DNAJC13 expression and autophagy flux

Knocking down DNAJC13 attenuates autophagic flux, whereas its overexpression promotes [14]. Accordingly, we examined the impact of RAB40 silencing on autophagic activity. RAB40C silencing alone did not affect DNAJC13 levels (data not shown); however, simultaneous RAB40A/B/C silencing resulted in significant increases in DNAJC13 expression (Fig. 5A and B and Suppl. Fig. 5A), which was accompanied by enhancement of LC3B trafficking to downstream acidic compartments (i.e., autolysosomes) (Fig. 5C). Collectively, these findings suggest that RAB40 indirectly regulates DNAJC13 degradation, thereby modulating autophagic flux.

Fig. 5.

RAB40 silencing significantly increases DNAJC13 levels and perturbs autophagic flux. (A, B) Triple silencing of RAB40A/B/C significantly increases DNAJC13 levels. Graphs show mean ± SEM. *p < 0.05 (Welch’s t-test; n = 5). (C) Autophagy flux assay using GFP-RFP-LC3B tandem sensor demonstrates that triple silencing of RAB40A/B/C significantly increases LC3B transport to acidic compartments (i.e., autolysosomes). Quantification of GFP/RFP ratios and the proportion of tandem fluorescent puncta is shown (mean ± SEM). *p < 0.05 (Welch’s t-test; n = 3). Scale bar: 20 μm.

3.6. Neddylation modulates RAB40C—CRL5-mediated DNAJC13 degradation

NEDD8 is a ubiquitin-like protein and neddylation (or NEDD8ylation) is the sequential process whereby NEDD8 is conjugated to its target proteins. The primary substrates of neddylation are the Cul family. CRL5 activation requires CUL5 neddylation, catalyzed by a neddylation enzyme cascade, consisting of the E1 NEDD8-activating enzyme (NAE1-UBA3), the E2 neddylation conjugating enzyme (UBE2F), and E3 neddylation ligase (RBX2) [31] (Fig. 6A). As shown in Fig. 6B and 6C and Suppl. Fig. 6B and 6C, a pull-down assay detected the CRL5 subunits NEDD8, RAB40C and RBX2, all required for CUL5-specific neddylation. Notably, a time-dependent DNAJC13 increase was observed in HEK293 cells exposed to MLN4924 (2.5 μM), a selective NAE1 inhibitor [32] (Fig. 6D and E, and Suppl. Fig. 6D). Conversely, UBE2F overexpression led to significant decrease in DNAJC13 levels (Fig. 6F and G, and Suppl. Fig. 6F). Consistent with the results obtained in HEK293 cells, MLN4924 treatment (1.25 μM for 12 h) showed a trend toward increased DNAJC13 expression in SH-SY5Y neuronal cells, although this did not reach statistical significance (Suppl. Fig. 10). Treatment with rotenone (0.25 μM for 12 h), a mitochondrial toxin used to generate animal models of PD, also showed a tendency to increase DNAJC13 expression. Combined treatment with rotenone and MLN4924 slightly increased DNAJC13 expression compared with rotenone treatment alone; however, the increase did not reach the level observed with MLN4924 treatment alone. Combined treatment with rotenone and MLN4924 substantially inhibited cell growth, suggesting that enhanced cytotoxicity may account for this limited increase.

Fig. 6.

CUL5 neddylation is essential for DNAJC13 degradation. (A) Schematic overview of the neddylation-dependent activation of CRL5. (B) Co-immunoprecipitation analysis detecting NEDD8 conjugation to myc-CUL5. (C) Co-immunoprecipitation assay showing RBX2 association with mCherry-RAB40C. (D, E) Treatment with MLN4924 (NAE1 inhibitor, 2.5 µM) induces a time-dependent increase in DNAJC13 levels. *p < 0.05, **p < 0.01, ***p < 0.001 (one-way ANOVA with Tukey’s post-hoc test; n = 3). (F, G) Conversely, UBE2F (E2 neddylation conjugating enzyme) overexpression reduces DNAJC13 levels. Graphs show mean ± SEM. *p < 0.05 (Welch’s t-test; n = 6).

4. Discussion

This study demonstrates that RAB40 regulates the degradation of DNAJC13, a protein linked to PARK21-linked familial PD. We identified RAB40A/B/C binding sites in both the N- and C-terminal regions of DNAJC13, including the IWN1 and IWN3 domains. DNAJC13 degradation primarily depends on the ubiquitin–proteasome system, and RAB40 expression enhances its ubiquitination. Conversely, RAB40 silencing elevates DNAJC13 levels with increased autophagic flux. RAB40C, highly expressed in midbrain dopaminergic neurons, forms a RAB40-CRL5 E3 ligase complex with CUL5, NEDD8, and RBX2. As CRL5 activation requires NEDD8-medated neddylation of CUL5, we found that inhibition of neddylation (MLN4924) increases DNAJC13 expression, whereas enhanced neddylation via UBE2F reduces it. Thus, neddylation acts upstream in RAB40-CRL5-mediated control of DNAJC13 proteostasis (Fig. 7).

Fig. 7.

Model for RAB40-dependent recruitment of DNAJC13 to CRL5 and its proteasomal degradation via neddylation. Neddylation is catalyzed by an E1 enzyme complex (NAE1-UBA3), one of two E2 NEDD8-conjugating enzymes (UBE2M or UBE2F), and substrate-specific NEDD8 E3 ligases. DNAJC13 bound to RAB40 is recruited to the activated CRL5 complex, thereby facilitating polyubiquitylation-mediated proteasomal degradation. This process is partly blocked by the NAE1 inhibitor, MLN4924.

Over a decade, multiple DNAJC co-chaperones (DNAJC5/CSPα, DNAJC6/Auxilin 1, DNAJC10/ERdj5, DNAJC12, DNAJC13, and DNAJC26/GAK) have been identified as causal molecules for familial forms of PD [33]. Among them, DNAJC13 (PARK21) has been most extensively studied since PARK21 patients exhibit core clinico-pathological features of sporadic PD [5]. Our previous study using cellular and fly models demonstrated that exogenous expression of the pathogenic variant DNAJC13 (DNAJC13^N855S) disrupted retrograde cargo transport from early endosomes, thereby causing locomotor dysfunction and dopaminergic cell loss with accumulation of insoluble αS [12]. More recently, DNAJC13 has been shown to play a key role in autophagy and trafficking of G protein coupled receptors (GPCRs) [[14], [15], [16]]. These results are interesting because αS can be phosphorylated by GPCR kinases, and hyper-phosphorylated, aggregated αS is degraded via macro-autophagy and chaperone-mediated autophagy [34,35].

Structurally, DNAJC13/RME-8 consists of a J-domain flanked by two IWN repeat domains on both sides. The J-domain binds the HSC70, thereby stimulates HSC70 ATPase activity for conformational and refolding capacities [36]. Our previous study in COS-1 cells demonstrated that WT DNAJC13 is localized to the perinuclear region, while a mutant lacking the N-terminal 453 aa (ΔN453) was distributed throughout the cytoplasm [10]. On the other hand, C-terminally-truncated mutants lacking the C-terminal 425 aa including the IWN4 (ΔC425) resulted in formation of large puncta and vacuoles. Kinetic analysis using reference molecules indicated that the phenotype induced by C-terminally-truncated mutant is due to perturbed trafficking through early endosomal compartments. In agreement with these findings, recent mutagenesis analysis using C. elegans coelomocytes revealed that the N- and C-terminal sequences of RME-8 have distinct functions [29]. Specifically, the N-terminal (IWN1 and IWN2) is important for endosomal recruitment while the C-terminal (IWN3 and IWN4) plays a role in endosomal microdomain localization and probably in substrate binding. In the present study, we investigated which region of DNAJC13 is necessary for binding to RAB40. In the N-terminal region, the ΔN453 mutant was insufficient to inhibit binding to RAB40s; a more widespread deletion lacking IWN1 (ΔN1083) was required. As for the C-terminal region, ΔC633 (lacking both IWN3 and IWN4), but not in ΔC425 (deleting only IWN4) failed to bind RAB40. According to the predicted structure of C. elegans RME-8 via AlphaFold (https://alphafold.ebi.ac.uk), much of the RME-8 structure consists of a series of five α-solenoids containing HEAT repeats with a short linker in between [29]. Remarkably, these linkers correspond to IWN domains and the J-domain forms a protruding structure adjacent to IWN2. Thus, the structural integrity of the region flanking the IWN-J-domain of DNAJC13 is important for binding to RAB40. We also examined the effect of the PD-linked N855S mutation on the interaction between DNAJC13 and RAB40 by co-immunoprecipitation. However, apparent reduction or loss of interaction was not observed (Suppl. Fig. 8). Since overexpression of N855S mutant DNAJC13 does not promote autophagy [14], the N855S mutation itself does not appear to impair the direct interaction between DNAJC13 and RAB40. Rather, it may contribute to disease pathogenesis by disrupting the intrinsic molecular function of DNAJC13.

Although the functional role of RAB40 subfamily proteins is not fully understood, they are involved in actin cytoskeletal regulation and cell migration, events possibly resulting from RAB40-CRL5-mediated ubiquitination/degradation of client proteins [[37], [38], [39], [40], [41]]. Evolutionarily, Rab40B and Rab40C are found in vertebrates whereas RAB40A and RAB40AL are solely expressed in other higher organisms than simian species [42]. In this experiment, no differences in binding affinity to DNAJC13 were observed among Rab40 paralogs. As expected, targeting only RAB40 was insufficient to increase DNAJC13 expression in HEK293 cells, and a triple knockdown of RAB40 subfamily was required. These results are consistent with previous findings showing that the RAB40 subfamily has functional redundancy in its interaction with CRL5 [40]. Supposedly, at the tissue/organ level, differences in RAB40 subfamily expression would have a greater impact on cellular homeostasis. In addition, we showed that the amino acid substitution of SOCS box in RAB40s (LPLP to AAAA) abrogated the interaction with CUL5, while binding to DNAJC13 did not change. These observations are consistent with previous observations showing that two switch regions located in the N-terminus (opposed to SOCS box) are important for RAB40 GTPase activity and binding to its client proteins [40,43]. Some RAB40 clients are not degraded through poly-ubiquitination, but instead undergo mono-ubiquitination, which alters their subcellular localization [40,43]. However, the following results: (i) DNAJC13 degradation mediated by UPS rather than autophagy, (ii) cellular DNAJC13 level regulated by RAB40 expression, and (iii) DNAJC13-RAB40 detection in the CRL5 complex, strongly support that Rab40-CRL5 plays a regulatory role in DNAJC13 degradation.

In the present study, we also explored the effect of RAB40 on the cellular function of DNAJC13. The pathogenesis of PARK21, caused by DNAJC13 missense mutations, has been proposed to involve haploinsufficiency [44]. Our result that reduced RAB40 expression suppresses DNAJC13 degradation and concomitantly enhances autophagic flux is consistent with recent findings indicating that DNAJC13 is a positive regulator of autophagic degradation [14]. We previously confirmed that PD-linked pathogenic mutant DNAJC13 disrupts cargo trafficking from the early endosome to the late and/or recycling endosome [12]. In line with this, a recent study has demonstrated that DNAJC13 controls surface retrieval as well as endo-lysosomal degradation of the δ-opioid receptor (DOR), a type of GPCR, and DNAJC13 depletion causes expansion of DOR-positive endosomes [16]. Presumably, these findings indicate that DNAJC13 is a dual-function protein promoting endosomal cargo sorting into both degradative and recycling pathways out of the endosomes.

5. Limitations

We acknowledge that there are several limitations in this study. First, the most experiments were conducted using non-neuronal cell lines. In the future, studies on neuronal lineage cells, preferably primary cultured neurons, are necessary. Second, to further validate and extend the findings of this study, analyses using animal models—specifically, studies using RAB40 knockout mice or RAB40 overexpression models, as well as evaluation of DNAJC13 expression levels following MLN4924 administration—would be important.

6. Conclusions

Our findings uncover a novel role of RAB40 as regulator of CRL5-UPS mediated DNAJC13 degradation. Increasing evidence has suggested that DNAJC13 is not only a causative molecule of familial PD, but is also involved in broader biological phenomena such as autophagy and the regulation of receptor expression on the cell surface [[14], [15], [16]]. Further elucidation of the regulatory mechanisms controlling DNAJC13 expression is expected to clarify its physiological and pathological significance.

Abbreviations

Ab: antibody

ANOVA: analysis of variance

αS: α-synuclein

Baf A1: Bafilomycin A1

CHX: Cycloheximide

Co-IP: Co-immunoprecipitation

CRL: Cullin-RING E3 ubiquitin ligase

CUL: Cullin protein

DANeu: dopaminergic neurons

DMEM: Dulbecco's modified Eagle medium

DNAJC13: DnaJ homolog C, member 13

DOR:δ-opioid receptor

EGFP: enhanced green fluorescencent protein

FAM21: family with sequence similarity 21

FBS: fetal bovine serum

GPCR: G protein coupled receptor

GWAS: genome-wide association study

HEK293: human embryonic kidney 293

HSP: heat shock protein

LB: Lewy body

NEDD8: neural precursor cell expressed developmentally downregulated 8 nTPM: normalized transcripts per million

PD: Parkinson’s disease

RME-8: receptor-mediated endocytosis-8 scRNA-seq: single-cell RNA sequencing

SDS-PAGE: sodium dodecyl sulfate-polyacrylamide gel electrophoresis siRNA: small interfering RNA

SNX: sorting nexin

SOCS: suppressor of cytokine signaling

TBST: Tris-buffered saline with 0.1% Tween 20

Ub: ubiquitin

UPS: ubiquitin-proteasome system

VPS: vacuolar protein sorting

WASH: Wiskott–Aldrich syndrome protein and SCAR homolog

WT: wild type

WB: Western blot

Consent for publication

Each author has read and approved the final manuscript version for submission.

Declaration of generative AI

The authors declare that no generative AI technologies were used in the preparation of this manuscript. Specifically, no such tools were employed for data acquisition, figure generation, or writing/editing of the text. All parts of the work were conducted by the authors, who take full responsibility for the content.

Ethics approval and consent to participate

Not applicable.

Funding

Funding for this project was supported in part by a Grant-in-Aid for Scientific Research (C) [grant number 20K07862 and 24K10614 (to N.S.); grant number 20K07896 and 23K06823 (to T.H.)], a Grant-in-Aid for Young Scientists [grant number 23K14769 (to S.Y.)], a Grant-in-Aid for Young Scientists (Start-up) [grant number 23K19557] (to S.I.) from the Ministry of Education, Culture, Sports, Science and Technology in Japan.

Suppl. Fig. 8 Parkinson’s disease-linked N855S mutation does not affect the interaction between DNAJC13 and RAB40C. Co-immunoprecipitation analysis was performed using HEK293 cells expressing EGFP-tagged WT or N855S mutant DNAJC13 together with mCherry-RAB40C. No apparent difference in binding to RAB40C was observed between WT and N855S mutant DNAJC13. Representative Western blots from three independent experiments are presented.

Suppl. Fig. 9 RNA expression pattern of the RAB40 subfamily across human central nervous system. RAB40A shows relatively high expression in the basal ganglia, RAB40AL in the cerebellum, RAB40B in the basal ganglia and spinal cord, and RAB40C in the cerebral cortex, basal ganglia, and cerebellum. Expression levels are expressed as normalized transcripts per million (nTPM), calculated by integrating RNA-seq data from the GTEx Consortium and HPA datasets with normalization across tissues.

Suppl. Fig 10 Changes in DNAJC13 expression in SH-SY5Y cells exposed to MLN4924 and/or rotenone. SH-SY5Y cells were treated with MLN4924 (1.25 μM) for 12 h in the presence or absence of rotenone (0.25 μM), and DNAJC13 expression levels were compared by Western blotting and densitometric analysis. Graphs show the mean ± SEM. *p < 0.05 (two-way ANOVA with Tukey’s post-hoc test; n = 3).

CRediT authorship contribution statement

Kazuki Sato: Writing – original draft, Visualization, Validation, Investigation, Formal analysis. Naoto Sugeno: Writing – review & editing, Visualization, Validation, Software, Investigation, Funding acquisition, Formal analysis, Data curation. Takafumi Hasegawa: Writing – review & editing, Visualization, Validation, Resources, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Shun Yoshida: Writing – review & editing, Resources, Investigation, Funding acquisition. Mitsunori Fukuda: Writing – review & editing, Resources, Formal analysis. Akemi Fujibayashi: Writing – review & editing, Resources, Formal analysis. Kiyotoshi Sekiguchi: Writing – review & editing, Resources, Formal analysis. Shun Ishiyama: Writing – review & editing, Investigation, Funding acquisition. Takaaki Nakamura: Writing – review & editing, Investigation, Funding acquisition. Ako Miyata: Writing – review & editing, Investigation. Satoshi Watanabe: Writing – review & editing, Investigation. Kensho Ikeda: Writing – review & editing, Investigation. Atsushi Takeda: Writing – review & editing, Supervision. Masashi Aoki: Writing – review & editing, Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.