FIELD OF INVENTION

The invention relates to the field of enzyme activation, in particular, the activation of ATP13A2. Described herein are methods of screening for compounds involved in activation of ATP13A2, based on functional properties such as the phosphorylation, localization, and binding properties of ATP13A2. Compounds which act as activators may be identified using ATP13A2 in assays using either full-length ATP13A2 or fragments, including N-terminal fragments.

BACKGROUND

Kufor-Rakeb syndrome (KRS) is a severe, juvenile-onset, autosomal recessive form of Parkinson's Disease (PD) associated with dementia. This atypical parkinsonism is caused by homozygous or compound heterozygous mutations in the ATP13A2 gene, which is harboured in the PARK9 (PD) susceptibility locus on chromosome lp36 [Ramirez A et al. (2006) Nat Genet. 38(10) : 1184-91]

In dogs and mice, loss of ATP13A2 function triggers neuronal ceroid lipofuscinosis (NCL), a lysosomal storage disorder, characterized by the accumulation of autofluorescent lipopigment [Farias FH et al Neurobiol Dis. (2011) 42(3) :468-74. ; Schultheis PJ et al. Hum Mol Genet. (2013) 22(10) :2067-82. Wohlke A et al PLoS Genet. 2011 Oct; 7(10) :el002304.] . Van de Warrenburg (2016) Eur J Hum Genet. 24(10) : 1460-1466 highlight that mutations in ATP13A2 might also be implicated in the onset of hereditary spastic paraplegia.

Yeast cells lacking the yeast PARK9 orthologue (Ypk9) develop a hypersensitivity towards heavy metals such as Mn ²⁺ , Zn ²⁺ , Cd ²⁺ , Ni ²⁺ and Se ²⁺ [Gitler AD, et al. Nat

Genet. 2009 Mar;41(3) :308-15; Tsunemi T & Krainc D. Hum Mol Genet. 2013 Dec 13. ;

Schmidt K et al. Biochem Biophys Res Commun. 2009 May 29;383(2) : 198-202.] . Also

NLF neuroblastoma cells overexpressing ATP13A2 are more resistant to Mn ²⁺ - and

Ni ²⁺ -induced cytotoxicity, whereas knock-down of ATP13A2 in SHSY5Y neuroblastoma cells are more sensitive toward Zn ²⁺ , but strangely not Mn ²⁺ [Covy JP, et al. J Neurosci

Res. 2012 Dec;90( 12) :2306-16; Kong SM et al., Hum Mol Genet. 2014 , 23(11) : 2816-

33]

In yeast, overexpression of YPK9 protects against a-synuclein toxicity, which was not observed with a functionally impaired mutant. The protective effect of ATP13A2 overexpression on a-synuclein toxicity is also conserved in Caenorhabditis elegans as well as in primary rat neurons [Gitler et al. Nat Genet. 2009 Mar;41(3) :308-15] . This links two genetic risk factors of PD, i.e. ATP13A2 and a-synuclein to environmental risk factors for PD, i.e. Mn ²⁺ and Zn ²⁺ , highlighting ATP13A2 as a critical mediator of PD. This also leads to the hypothesis that increasing the activity of ATP13A2 in neurons, e.g. by overexpression of ATP13A2, would offer protection against α-synuclein toxicity, providing an interesting potential therapeutic strategy to treat PD, NCL or other a- synucleinopathies.

The protective effect of ATP13A2 seems to relate to its effect on lysosomal function. ATP13A2 is targeted to the late endosomal/lysosomal compartments in mammalian cells. Neuronal fitness depends on optimal lysosomal function and efficient lysosomal delivery of proteins and organelles for subsequent breakdown [ Dehay B, et al . Mov Disord. 2013 Jun;28(6) :725-32; Usenovic M & Krainc D. Autophagy. 2012 Jun;8(6) :987-8.] . Thus a role for ATP13A2 in lysosomal function and autophagy regulation has been proposed [Dehay B, et al. ATP13A2 gets into the groove. Autophagy. 2012 ;8(9) : 1389-91; Gusdon AM et al. Neurobiol Dis. 2012 ;45(3) :962- 72] .

Mutations in or knockdown of ATP13A2 lead to lysosomal dysfunctions, including impaired lysosomal acidification, decreased degradation of lysosomal substrates, decreased lysosomal-mediated clearance of autophagosomes and impaired proteolytic processing of lysosomal enzymes [Dehay B et al. Proc Natl Acad Sci USA. 2012 12; 109(24) :9611-6] . Furthermore, recent data implicate loss of ATP 13 A2 function to impaired mitochondrial clearance and oxidative stress, lending support to the importance of lysosomes for mitochondrial quality control [Xu Q et al. Neurochem. 2012 122(2) :251-259.; Grunewald A et al. Neurobiol Aging. 2012;33(8) : 1843 el-7; Park JS, et al. Hum Mol Genet. 2014 Jan 15].

ATP13A2 belongs to the superfamily of P-type transport ATPases. P-type ATPases are membrane proteins, which use the energy of ATP hydrolysis to actively transport substrates across the membrane or between membrane leaflets [Kuhlbrandt W. Nat Rev Mol Cell Biol. 2004 Apr; 5(4) : 282-95] . During the transport cycle, a phospho- intermediate is formed on a critically conserved aspartate residue. The associated phosphorylation and dephosphorylation events are tightly coupled to ATP hydrolysis, substrate binding and transport. All P-type ATPases switch between an El and E2 conformation to alternate the access to the transport binding sites from both sides of the membrane (Fig. IB). During this cycle an aspartyl-phospho-intermediate is formed on the conserved aspartate of the consensus motif DKTG[T/S]. P-type ATPases belonging to the Pl-3 subfamilies transport inorganic ions across the membrane, generating crucial ion gradients required for vital cellular processes. The P-type ATPases that belong to the P4 subfamily, flip lipids from one membrane leaflet to the other and are termed flippases. Flippases generate lipid asymmetry that is important for vesicle budding/fusion or for exposing/removing lipids in one membrane leaflet to exert a signaling function [Palmgren MG, et al. Annu Rev Biophys. 2011 ;40 :243-66; Sorensen DM . Biochim Biophys Acta. 2010; 1797 :846-855] .

ATP13A2 belongs to the five human isoforms of the P5-type ATPases (ATP13A1-5) for which the transport substrate remains unidentified [Ramirez A et al. cited above; Kuhlbrandt W. Nat Rev Mol Cell Biol. 2004 5(4) : 282-295.] . P-type ATPases are multi- span transmembrane proteins, comprising a common stretch of six conserved membrane helices (M l-6) that form the substrate binding sites and entrance/exit pathways. Most P-type ATPases harbor at least four additional membrane helices at the C-terminal end (M7-10). Interestingly, topology predictions indicate that ATP13A2 and other P5 ATPases may contain at least one additional N-terminal M helix [Sorensen DM et al. Biochim Biophys Acta. 2010 1797(6-7) :846-855.] . Extra N-terminal M stretches are uncommon in P-type ATPases and are only present in the heavy metal pumps belonging to the PIB-type ATPases. Like for the heavy metal pumps, the N- terminal helices in the P5-type ATPases might be important for subcellular targeting, substrate recognition, delivery, transport and/or regulation [Gourdon P et al. Nature. 2011 475 : 59-64.]

Regulated turnover of intracellular proteins maintains the quality of the proteome (proteostasis), which is required for a cell to function properly. In line, proteostasis has been shown to be disturbed during aging as well as in many pathological conditions, including neurodegeneration [Tanaka, K. and N. Matsuda Biochim Biophys Acta, 2014. 1843(1) : p. 197-204.] . To maintain proteostasis and prevent proteotoxicity as a result of the accumulation of aged or aberrantly folded proteins, cells depend on two key proteolytic systems; the proteasome and the endo-/lysosomal system. Non-functional proteins are usually tagged with various ubiquitin (Ub) signals to be delivered to the 26S proteasome or to the lysosomes through different autophagy pathways. The latter can involve; i) sequestration of damaged proteins/aggregates within autophagosomes (macroautophagy or simply autophagy) that then fuse with lysosomes thereby delivering their contents for degradation and recycling ; ii) the formation of vesicles directly th rough the invagination of the lysosomal membrane followed by the degradation of the proteins in the lysosomal lumen by hydrolytic enzymes (microautophagy) ; and iii) pathways involving chaperone-med iated recognition of soluble proteins and their trafficking across the lysosomal membrane th rough a receptor-mediated mechanism (chaperone-mediated autophagy or CMA) or directly to the late endosomes (endo-/lysosomal microautophagy) [Baixauli, F. et al . Front Immunol, 2014. 5 : p. 403] .

Moreover, autophagosomes can enter the endocytic pathway at multiple stages, including early endosome (EE), multivesicular bodies (MVBs) and late endosomes (LEs) as well as fusing with lysosomes, thus highlighting the dynamic complexity of the endo-/lysosomal degradation pathway [Eskelinen, E. L., et al . Autophagy, 2011. 7(9) : p. 935-56; Repnik, U . et al . Cold Spring Harb Perspect Biol, 2013. 5( 1) : p. a008755] . The relevance of these quality control mechanisms for cellular homeostasis is fu rther supported by the recognition that impairments in endo-/lysosomal degradation and sorting pathways underlie the aetiology of neurodegenerative disorders suggesting the importa nce of proteostasis in these diseases [Gestwicki, J .E. and D. Garza, Prog Mol Biol Tra nsl Sci, 2012. 107 : p. 327-53. ; Gao, X. a nd H. Hu, Acta Biochim Biophys Sin (Shanghai), 2008. 40(7) : p. 612-8.] . The accumulation of misfolded alpha-synuclein aggregates is a hallmark of PD and can be a direct consequence of impaired proteasomal or lysosomal function . Also, alpha-synuclein is removed via both the proteasome and autophagy dependent lysosomal pathways.

ATP13A2/PARK9 is a P5 type ATPase consisting of 10 transmembrane domains, with a N- and C-terminus exposed into the cytoplasm [Holemans, T., et al.. Proc Natl Acad Sci USA, 2015.112: p. 9040-5] To date the function of ATP13A2 is still elusive, yet it is known to reside in the late endo-/lysosomes and to be involved in key cell biological functions [van Veen S, et al. Front Mol Neurosci. 2014;7 :48.] Loss of function or mutations in ATP13A2 underlie Kufor-Rakeb syndrome, a form of autosomal Pa rkinsonism [Ramirez et al . cited above] and several studies have highlighted a relevant protective role of ATP13A2 in the suppression of a-synuclein toxicity [Gitler, A. D., et al . Nat Genet, 2009. 41(3) : p. 308- 15; Usenovic, M ., et al. J Neurosci, 2012. 32( 12) : p. 4240-6. ; Usenovic, M . and D. Krainc, Autophagy, 2012. 8(6) : p. 987-8.] . Recent studies have shown that ATP13A2 is required for the efficient functionality of lysosomes and autophagy [Usenovic in Autophagy, cited above] and for the biogenesis and secretion of exosomes [Tsunemi, T. et al 2014. J Neurosci. 34(46) : 15281- 15287.]. However the molecular mechanisms by which ATP13A2 exerts this regulatory capacity remains unclear. SUMMARY OF INVENTION

The present invention describes ATP13A2, a late endo-/lysosomal P5-type transport ATPase important for lysosomal functionality, but for which the transported substrate remains unknown up to present.

ATP13A2 loss of function (LOF) mutations lead to diseases such as Kufor Rakeb syndrome, early onset Parkinson's disease, neuronal ceroid lipofuscinosis and heriditary spastic paraplegia. This spectrum of disorders is marked by impaired lysosomal (lyso) and mitochondrial (mito) function. Hence, improving ATP13A2 activity is candidate therapeutic strategy in a spectrum of related neurodegenerative disorders that are marked by impaired lyso/mito function.

Data demonstrated that ATP13A2 expression modulates endocytic trafficking and sorting of SHSY5Y cells under proteotoxic stress conditions. For the first time we show that ATP13A2 incites endocytic trafficking/sorting through its cytosolic N terminal domain. Furthermore we highlight for the very first time mechanistic insights whereby ATP13A2 possesses a significant scaffolding function driven by its N terminal domain, requires the interaction with the signalling lipid PI(3,5)P2 to potentiate the release of nano-vesicles which buffers intracellular proteotoxic stress insisted by proteasomal inhibition.

It is shown that ATP13A2 provides cellular protection in various Parkinson's disease models such as alpha-synuclein toxicity, mitochondrial toxicity due to complex I inhibition, and metal toxicity (such as Fe, Mn, Zn).

It has been demonstrated that ATP13A2 resides in an inactive state (E1P) and that activation of ATP13A2 depends on the binding of two signalling lipids, phosphatidic acid (PA) and phosphatidylinositol(3,5)bisphosphate (PI(3,5)P2) that interact with the N- terminus of ATP13A2. This indicates that the N-terminus exerts an auto-inhibitory function. Targeting the auto-inhibition mechanism is a candidate strategy to activate ATP13A2.

It has further been demonstrated that increased ATP13A2 levels are detected in surviving neurons of Parkinson's disease brains, and that ATP13A2 dysfunction leads to lysosomal deficiency. The present invention demonstrated a link between ATP13A2 function and risk factors of Parkinson's disease. It has been further demonstrated that ATP13A2 overexpression protects against a-synuclein toxicity, and that ATP13A2 deletion sensitizes cells to Mn ²⁺ and Zn ²⁺ .

Mutations and Knock-Downs of ATP13A2 lead to lysosomal dysfunction with reduced lysosomal acidification, decreased degradation of lysosomal substrates, impaired autophagy, and accumulation of fragmented mitochondria.

These observations point to a central role of ATP13A2 in Parkinson's disease and indicates that upregulation of ATP13A2 activity and lysosomal degradation capacity has a therapeutic potential in the context of Parkinson's disease.

ATP13A2 seems to be in an inactive state, waiting for further activation or binding of its substrate. Initially, ATP13A2 was proposed to be a lysosomal Mn or Zn transporter, although no direct proof was provided. However, using the phosphorylation assay it was observed that ATP13A2 phospho-enzyme levels are not affected by Mn or Zn. The N-terminal interacting lipids PA and PI(3,5)P2 stimulate phospho-enzyme formation, pointing to a critical role of these lipids on the catalytic cycle of ATP13A2.

Herein disclosed is a Parkinson's disease in vitro cell model of dopaminergic neurons in Parkinson's disease. Herein to mimic the pathophysiology of Parkinson's disease, rotenone is used. A PIKfyve inhibitor is used to block the production of PI(3,5)P2. A Phospoholipase D (PLD) inhibitor is used to block production of PA.

The present invention demonstrates that ATP13A2 protects against mitochondrial stress via a novel lysosomal dependent clearance pathway.

Herein are stable cell lines use for screening purposes such as neuroblastoma cell lines stably overexpressing ATP13A2 or a shortened version thereof. In addition viral vectors of the N-terminal fragments of ATP13A2 or full length ATP13A2 are used for stereotactic injections into the brain to promote local expression of ATP13A2 in dopaminergic neurons.

In this model it has been demonstrated that ATP13A2 protects cells against a panel of Parkinson's disease associated agents, ranging from an array of heavy metals to mitochondrial targeting specific toxins such as rotenone, MPP+, 6-OHDA and Rhodamine 6G, Rotenone or zinc alone or in combination with Bafilomycin.

The autophosphorylation data show that rotenone treatment activates the autophosphorylation of ATP13A2, in line with the PA/PI(3,5)P2 mediated binding to the N-terminus that is required for the protective effect during rotenone exposure. The cell models further show that ATP13A2 protects against a spectrum of reactive oxygen species (ROS) inducing agents such as peroxide and superperoxide.

The production and/or accumulation of ROS mainly occurs at the level of the mitochondria. More importantly, ATP13A2 minimizes the levels of ROS that are exposed to the cell at any one time. As mitochondrial defects can be found in a spectrum of diseases, targeting the activation of a protein capable of minimizing this unwanted attribute is beneficial.

Due to the significant correlation between ATP13A2, ROS inducing agents, mitochondria and cell death the importance of ROS was investigated. The protective phenotype of ATP13A2 overexpression could be mimicked in the ATP13A2 knockdown cells by the addition of an anti-oxidant (such as NAC, VIT C, URIC Acid, Mito-TEMPO). The most significant outcomes were observed when the anti-oxidant was targeted specifically to the mitochondria, highlighting mitochondrial dysfunction as the main source of cell stress and death.

It was demonstrated that ATP13A2 protects at the mitochondrial level against mitochondrial dysfunction, and maintains a pool of healthy mitochondria that are able to function. This occurs with various insults like rotenone, zinc, but also for other metals or complex I inhibitors the same phenotype can be observed. This points to the direct connection between ATP13A2 and mitochondrial protection.

The effect of ATP13A2 on mitochondrial mass, together with the lysosomal dependency of the phenotype, shows that ATP13A2 modulates mitochondrial clearance, and this independent of mitophagy, and that a novel mitochondrial clearance pathway is involved, which is regulated by ATP13A2. ATP13A2 minimizes rotenone induced mtUPR (mitochondrial Unfolded Protein Response). mtUPR induction is used as a marker for mitochondrial dysfunction and protein stress in the mitochondria. The data demonstrate that the success of ATP13A2 in maintaining mitochondrial functionality results in the down regulation of a key stress response, mtUPR. This response initially incites pro-survival attributes, such as the upregulation of proteases, chaperones(HSP60) and transcription factors (CHOP and CEBP/Beta), but also the translocation of the transcription factor CEBP/Beta to the nucleus.

ATP13A2 mutations recapitulate the mitochondrial phenotype in patient derived fibroblasts. Basal mitochondrial functionality/health of two independent patient derived fibroblast models of ATP13A2, possessing either the T512I or DeITT truncated variants were assessed, in comparison to wild-type controls. Models were assessed for mitochondrial potential. Investigating the mitochondrial phenotype of ATP13A2 in patient derived fibroblasts confirmed the impact of ATP13A2 on mitochondrial homeostasis (mtUPR) and clearance (mito mass) . Importantly, differences on a basal level highlight the importance of functional ATP13A2 in cell functionality and mitochondrial-associated disease prevention.

ATP13A2 patient derived fibroblasts display heightened ROS dependent sensitivity to rotenone induced mitochondrial stress. This indicates that the interplay between genetic and environmental risk factors of disease is crucial in the outcome of cell survival and highlights the importance of ATP13A2 in cellular homeostasis

ATP13A2 promotes mitochondrial homeostasis independent of autophagy, which was assessed by western blot analysis for LC3 and p62, and by fluorescence microscopy evaluating the two colour autophagy probe GFP-RFP-LC3. Note that mitophagy is fully functional upon exposure to CCCP and that ATP13A2 has no effect on mitophagy. We therefore conclude that the effect of ATP13A2 on mitochondrial mass occurs independently of mitophagy. The ATP13A2-mediated protective effect at the mitochondria requires lysosomal functionality, since inhibition of lysosomes impairs the mitochondrial clearance phenotype. Together, our data highlight the possibility of a novel lysosomal clearance pathway that utilizes the existing endocytic potential of the cell. Therefore modalities to incite this functionality may be important for a number of disorders that demonstrate mitochondrial defects.

The aforementioned phenotype is dependent on the functionality of the lysosomes And ATP13A2 promotes mitochondrial-lysosomal co-localization in a lysosomal functionality dependent manner. Therefore as functional ATP13A2 also mediates the health of the cellular lysosomal pool, the activation of ATP13A2 would work on multiple fronts to promote cellular homeostasis and the prevention of cell death/loss. It could work directly to mediate mitochondria health and also to promote global lysosomal functionality and increase removal of unwanted proteins/aggregates such as alpha synuclein.

ATP13A2 promotes mitochondrial-late endo-/lysosomal contacts indicating that ATP13A2 can interact directly at the level of the mitochondria. Here we show the presence of ATP13A2 in a a cellular fraction containing the mitochondria, which can be increased by the addition of the mitochondrial toxin rotenone. Interestingly, this does not correlate with the lysosomal marker LAMP1. This strongly supports the view that ATP13A2 promotes mito/lyso contacts, which might facilitate the transfer of damaged proteins/material to the lysosome/late endosome for clearance via exosomes and/or degradation into the lysosomes.

Dimerization of ATP13A2 occurs via intermolecular interaction of the N-terminus with most likely a binding pocket in the cytosolic domains (based on sequence analysis we predict that the interaction site for the N-terminus might be found in the P5-type ATPase specific insert sequence in the cytosolic loop between transmembrane regions M4-5) . The dimer represents the inactive form of the protein. Promoting the dimer to monomer transition leads:

To ATP13A2 transport activation, which protects neurons against mitochondrial and metal toxicity stress, which is implicated in Parkinson's disease and related neurodegenerative disorders.

b) Promotes the scaffolding functionality of ATP13A2, i.e. Independent of the ATP13A2 transport activity, which leads to the clearance of ubiquitinated proteins and alpha- synuclein via exosomes in conditions of protein stress as occurs in Parkinson's disease. Thus, both mechanisms offer possibilities to improve neuroprotection in Parkinson's disease and related neurodegenerative disorders.

To relieve autoinhibition interaction of PA and PI(3,5)P2 at the N-terminus of ATP13A2 is required. This will induce monomerisation and hence, activation.

The dimerization status of ATP13A2 can be used as a read-out to assess the functional status of ATP13A2 in Parkinson's disease mutants.

In the dimer conformation, the N-terminus is buried and inaccessible for interacting proteins. Monomerization, removal of the N-terminal interaction site in ATP13A2 or prevention of the N-terminal interaction with its binding site will make the N-terminus accessible to putative interactors (e.g. a-synuclein) and stimulate the clearance phenotype.

Multiple positively charged residues in previously identified lipid binding sites (LBS) were substituted for Ala (Holemans et al, PNAS 2015, cited above). LBS3 (KRVLR) [SEQ ID NO: 5] overlaps with an alternative splicing site rendering an insertion of five additional residues in splice variant 1 of ATP13A2. Such LBS mutants occur predominantly in the dimeric state. These dimers are very stable as they are resistant to SDS treatment and only partially affected by boiling and urea.

The LBS mutants are unable to auto-phosphorylate, showing that these mutant proteins are catalytically inactive, possibly by a strong auto-inhibition by the N- terminus. Mutation of LBS2 (RLRLR) [SEQ ID NO : 4] and LBS3 (KRVLR) [SEQ ID NO: 5] perturb the ability of ATP13A2 to protect against mitochondrial stress. These mutants behave similar to knock-down of ATP13A2 indicating that these mutants have a dominant negative effect on the endogenously expressed ATP13A2 most likely by forming a stable inactive dimer. The dimerization model might explain the observed dominant negative phenotype : overexpressed mutant ATP13A2 will also form dimers with endogenous ATP13A2 leading to inactivation of the endogenous protein.

Only for the peptides with LBS2/3 mutations we observed a stable interaction with full- length ATP13A2 on a denaturing gel. This interaction must be very stable as it is resistant to SDS-PAGE. At the same time this means that, although there is no interaction with the WT peptide, the interaction might be weak and lost under denaturing conditions. Since N-terminal peptides with mutations in LBS2 or LBS3, bind very strongly to ATP13A2, this supports the dimerization model in which the N- terminus of one monomer interacts with the full length protein of the other monomer, and that the lipid binding sites regulate this interaction which impacts on activity.

Low amounts of WT N-terminal peptide increase auto-phosphorylation of ATP13A2, whereas high amounts of WT N-terminal peptide do not affect auto-phosphorylation of ATP13A2. As the WT N-terminal peptide is able to influence ATP13A2 phospho-enzyme formation, the peptide must interact with ATP13A2, although not visible on SDS-PAGE. The data are in line with the dimerization model since here the N-terminal peptide leads to increased ATP13A2 autophosphorylation activity, most likely by out-competing the interaction of the N-terminus with the binding site for auto-inhibition.

Whereas a WT peptide stimulates the activity of ATP13A2, the N-terminal ATP13A2 peptides with mutations in LBS2 or LBS3 strongly inhibit ATP13A2 auto- phosphorylation. The dominant negative phenotype of LBS2/3 may result from a direct and stable interaction on an inhibitory binding site for the N-terminus on ATP13A2, further confirming the dimerization model, which is regulated by the lipid binding site domains located in the N-terminus.

An antibody against a peptide of the N-terminus of ATP13A2 (PSPQSQAEDGRSQAAVGAV) [SEQ ID NO:6] was able to increase phospho-enzyme formation of ATP13A2, presumably by preventing the binding of the N-terminus to the auto-inhibitory binding site.

Stable cell lines with overexpression/knockdown of ATP13A2 have been generated and the late endo/lysosomal targeting of ATP13A2 was confirmed in these cells. Inhibition of the proteasome with bortezomib leads to accumulation of ubiquitinated proteins. ATP13A2 modulates the level of ubiquitinated proteins: More ATP13A2 reduces the ubiquitinated proteins, whereas knockdown has the opposite effect.

Importantly, a catalytic dead mutant D508N promotes the removal of ubiquitinated proteins to the same extent as WT protein, indicating that this phenotype is independent of ATP13A2 activity. We refer to this new phenotype as the scaffold function of ATP13A2. Even more strikingly, an N-terminal fragment of ATP13A2 corresponding to the first 251 amino acids also promotes Ub clearance, suggesting that the scaffold function highly depends on the N-terminus of ATP13A2. The N-terminus might be considered as a clearance tool for ubiquitinated proteins.

In a SHSY5Y cell model for Parkinson's disease with overexpression of alpha-synuclein, it is shown that not only WT, but also disease or catalytic dead mutants are equally potent in removing alpha-synuclein via exosomes.

Thus, the alpha-syn clearance and exosome production occurs independent of ATP13A2 transport activity. As the N-terminal fragment is unable to form dimers it therefore displays a constitutively activated scaffold function.

Thus, the disruption of dimers into monomers can also promote alpha-syn removal via exosomes.

The activity of P-type ATPases can be assessed by measuring ATP hydrolysis. Samples are prepared with purified protein and ATP and left for 30 min for ATP hydrolysis to occur. Molybdate is added which will form a complex with Pi released during ATP hydrolysis. This complex is reduced by ascorbic acid to form the blue coloured β-keggin ion. Arsenite is added to complex the remaining molybdate. The amount of the blue coloured ion produced is proportional to the a mount of Pi present and the absorption can be measured using a colorimeter to determine the amount of phosphorus. We will screen for activating compounds by following an increase in ATPase activity. Note that any other detection method to measure ADP or Pi production can be used, such as a luciferase based system to monitor the ADP production, which is converted by ATP by an enzymatic reaction, which then stimulates luciferase (light emission) activity.

Herein disclosed are the following statements:

1. A method for detecting a compound that targets the auto-inhibitory mechanism of ATP13A2 protein or ATP13A1/3-5 isoforms leading to a modulation of ATPase activity, the method comprising the steps of: providing an assay system comprising ATP13A2 protein or ATP13A1/3-5 isoforms, and a source of ATP,

adding a test compound to said assay system, wherein the test compound is not a phosphatidic acid (PA) and/or phosphatidylinositol(3,5)bisphosphate (PI(3,5)P2), determining whether the test compound modifies ATP hydrolysis, compared to a control assay without test compound,

wherein a test compound which modifies ATP hydrolysis is identified as modulator of ATPase activity. In alternative embodiments autophosphorylation is used as read out for the activity of the protein.

The claimed method provides evidence for the fact that ATP13A2 protein or ATP13A1/3-5 isoforms can be activated by phospholipids, but also by other metabolites. The claimed method allows to screen for activators and inhibitors of ATP metabolisation, such as endogenous metabolites as well as synthetic organic compounds.

2. The method according to statement 1, wherein the test compound is not a phospholid.

3. The method according to statement 1 or 2, wherein ATPase activity is measured by a decrease in ATP and/or an increase in ADP and/or an increase in inorganic phosphate.

4. The method according to any one of statements 1 to 3, wherein the test compound is a sequence with SEQ ID 1, 2 or 3 or a peptide fragment thereof.

5. The method according to any one of statements 1 to 3, wherein the test compound is a monomer promoting fragment of ATP13A2 or ATP13A1/3-5.

6. The method according to any one of statements 1 to 3, wherein the test compound is a compound that specifically binds to a peptide with the amino acid sequence

PSPQSQAEDGRSQAAVGAV [SEQ ID NO: 6].

7. The method according to statement 6, wherein the peptide binding compound is an antibody, or antigen binding portion thereof, such as a nanobody.

8. The method according to any one of statements 1 to 7, wherein the ATP13A2 protein or ATP13A1/3-5 isoforms are purified recombinant proteins, such as obtained from yeast. ATP13A1-5, which are membrane proteins can be purified from yeast. In alternative embodiments, the protein is purified from microsomes whereby membranes components remain associated with the protein. 9. The method according to any one of statements 1 to 8, wherein the ATP13A2 protein or ATP13A1/3-5 isoforms are expressed in microsomes harvested from cells.

10. The method according to any one of statements 1 to 9, wherein ATP hydrolysis is determined with e.g. a colorimetric assay, an autophosphorylation assay a luminescence assay or a luciferase based assay.

11. The method according to any one of statements 1 to 10, wherein the assay is for identifying compounds, which increase ATPase activity.

12. The method according to any one of statements 1 to 11, wherein the assay is performed in the presence of a phospholipid such phosphatidic acid (PA) and/or phosphatidylinositol(3,5)bisphosphate (PI(3,5)P2) in order to identify compounds which inhibit or enhance the ATPase activity. In this way the

13. A protein binding compound binding to the N-terminal part of ATP13A2 and activating ATP13A2 for use in the treatment of neurodegenerative disorders that are marked by impaired lysosomal and or mitochondrial function.

14. The protein binding compound for use in the treatment according to statement 13, wherein the protein binding compound binds specifically to a peptide comprising the amino acid sequence PSPQSQAEDGRSQAAVGAV [SEQ ID NO:6] .

15. The protein binding compound according to statement 13 or 14, for use in the treatment according to statement 13, wherein the protein binding compound is an antibody, or antigen binding portion thereof.

16. The protein binding compound according to statement 13 or 14, for use in the treatment according to statement 13, wherein the protein binding compound is a nanobody.

17. The protein binding compound according to statement 16, for use in the treatment according to statement 13, wherein the antibody or nanobody has been modified for intracellular uptake, such as peptide signals.

18. The protein binding compound according to statement 16 or 17, for use in the treatment according to statement 13, wherein the nanobody is delivered as a gene transfer vector comprising a DNA sequence encoding the nanobody.

19. A method for detecting compounds with therapeutic activity against a neurodegenerative disorder that is marked by impaired lysosomal and or mitochondrial function, the method comprising the steps of:

providing an assay system with ATP13A2 protein or ATP13A1/3-5 isoforms or a monomer promoting fragment thereof, adding a test compound to said system,

determining the effect of the test compound on dimer formation,

wherein a test compound which impairs or inhibits dimer formation is indicative for therapeutic activity against said disorder.

20. The method according to statement 19, wherein the ATP13A2 or ATP13A1/3-5 isoform protein or a dimer forming fragment thereof comprises a reporter system for dimer formation.

21. The method according to statement 18, wherein the reporter system is a protein complementation assay.

22. The method according to statement 20, wherein the reporter system is a luciferase protein, of which enzymatic inactive fragments form a fusion protein with the ATP13A2 protein or ATP13A1/3-5 isoform or a dimer forming fragment thereof, whereby the upon dimer formation, the inactive fragments of luciferase become positioned in each other proximity and perform luciferase activity. This method can be equally performed with another protein complementation assay, based on the GPF constructs or Fret based assays.

23. The method according to statement 20, wherein the reporter system is a FRET system. In alternative methods dimerization is determined by e.g. FSEC chromatography, anisotropic methods, native electrophoresis systems.

24. The method according to any one of statements 19 to 23, wherein said neurodegenerative disorder is Parkinson's disease.

25. The method according to any one of statements 19 to 23, wherein said neurodegenerative disorder is neuronal ceroid lipofuscinosis.

26. The method according to any one of statements 19 to 23, wherein said neurodegenerative disorder is spastic paraplegia.

27. The method according to any one of statements 19 to 26, wherein the test compound is an analogue or mimic of PA or PI(3,5)P2.

28. The method according to any one of statements 19 to 27, wherein the test compound is a fragment of the N-terminal part of ATP13A2 with SEQ ID 1, 2 or 3, or of the N-terminal part of ATP13A1/3-5 isoforms.

29. The method according to any one of statements 19 to 28 wherein said ATP13A2 protein or ATP13A1/3-5 isoforms or a monomer promoting fragment thereof or is a recombinant protein expressed in yeast. In alternative embodiments, the protein is purified from microsomes whereby membranes components remain associated with the protein.

30. A protein selected from the group consisting of the Nterminal of ATP13A2 with SEQ ID 1, 2 or 3 or a peptide fragment thereof, or a corresponding sequence of ATP13A1,3- 5, for use in the treatment of neurodegenerative disorders that are marked by impaired lysosomal and or mitochondrial function.

31. The protein or peptide fragment according to statement 30 for use in the treatment according to statement 30, wherein the peptide interacts with the dimer formation of ATP13A2 and activates ATP13A2.

32. The peptide fragment according to statement 30 or 31, for use in the treatment according to statement 30, wherein the peptide fragment has a length of between 5 and 50 amino acids.

33. The peptide fragment according to any one of statements 30 to 32, for use in the treatment according to statement 30, wherein the peptide fragment comprises the sequence FRWK [SEQ ID NO :7], RLRLR [SEQ ID NO :4], KRVLR [SEQ ID NO: 5] PS PQSQ AE DG RSQ AAVG AV [SEQ ID NO: 6]

34. The peptide fragment according to any one of statements 28 to 33, for use in the treatment according to statement 30, wherein the peptide fragment comprises the sequence PSPQSQAEDGRSQAAVGAV [SEQ ID NO : 6].

35. The peptide fragment according to any one of statements 30 to 33, for use in the treatment according to statement 30, wherein the peptide fragment comprises the sequence FRWK [SEQ ID NO: 7] and FRWKPLWGVRLRLR [SEQ ID NO:8] .

36. The peptide fragment according to any one of statements 30 to 35, for use in the treatment according to statement 30 wherein the peptide fragment further comprises a signal sequence for cellular uptake.

37. The peptide fragment according to any one of statements 30 to 36, for use in the treatment according to statement 30, wherein the peptide is modified such as by the introduction of modified or non-natural amino acids, modification of the N or C terminus, or by providing a peptidomic of said peptide.

38. A method for detecting compounds with therapeutic activity against a neurodegenerative disorder which is marked by impaired lysosomal and or mitochondrial function comprising the steps of: providing an assay system comprising mammalian cells which are experiencing mitochondrial or proteasome stress, and wherein one of ATP13A1 to 5 is overexpressed, knocked down or knocked-out,

adding a test compound to said system,

determining whether the test compound increases the interaction or overlap between lysosomes and mitochondria,

wherein a test compound which promotes interaction or overlap is indicative for therapeutic activity against said disorder.

When this assay is performed with a control wherein the wild type of the corresponding protein is expressed at wild-type levels, it is possible to determine whether the determine change of interaction or overlap between lysosomes and mitochondria is ATP13A dependent.

39. The method according to statement 38, wherein as an alternative the method comprises the step of testing mitochondrial functionality and mass, lysosomal mass/distribution, lysosomal functionality or cell survival and cell death.

40. The method according to statement 38 or 39, wherein said neurodegenerative disorder is Parkinson's disease, neuronal ceroid lipofuscinosis or spastic paraplegia.

41. The method according to statement 40, wherein said mammalian cells which are experiencing mitochondrial stress are cells from Parkinson's disease patients.

42. A method for detecting fragments and/or mutants of the N-terminal part of ATP13A2, or related isoforms ATP13A1/3-5, with therapeutic activity in neurodegenerative disorders that are marked by impaired lysosomal and or mitochondrial function comprising the steps of:

providing an assay system with cells producing alpha-synuclein aggregates or toxicity, or accumulation of ubiquitinated proteins ,

determining the effect of a fragment and/or mutant of the N-terminal part of ATP13A2 on clearance of alpha-synuclein aggregates, ubiquitinated proteins or toxicity, wherein clearance of alpha-synuclein aggregates or ubiquitinated protein or reduction of alpha-synculein toxicity is indicative for therapeutic activity against said disorder. 43. The method according to statement 42, wherein the clearance of alpha synuclein aggregates is measured by a decrease in the overall ubiquitination in said cells.

44. The method according to statement 42, wherein the clearance of alpha synuclein aggregates is measured by an increase of exosomal proteins. 45. The method according to statement 42, wherein the clearance of alpha synuclein aggregates is measured by an increase in co-localisation of mitochondria and lysosomes or re-localization of lysosomes to the proximity of the plasma membrane.

46. The method according to any one of statements 42 to 45, wherein said neurodegenerative disorder is Parkinson's disease.

DETAILED DESCRIPTION OF INVENTION

Figure legends

Figure 1. ATP13A2 co-localizes with LAMP1 (endo-/lysosomal marker).

(A) Representative western blots demonstrating the ATP13A2 expression patterns in SHSY5Y neuroblastoma cell lines, cell lines stably transduced by lentiviral-based technology with the overexpression of WT ATP13A2, catalytic inactive mutant (D508N or DN) and sh-ATP13A2 (sh-ATP13A2, mRNA levels were reduced by 75.4 ± 8.9%) in comparison to firefly luciferase as a control (control).

(B) Representative confocal microscopy images of control (top), shATP13A2 knocked down (bottom) and ATP13A2 GFP overexpression (middle) in SHSY5Y (B) cell lines. Co- localization in merged images is revealed in the panels at the right. Data are representative of a minimum of three independent experiments.

Figure 2. Following Bortezomib (Borte) treatment, cells overexpressing ATP13A2 display a reduced accumulation of Ub-proteins independent of catalytic activity of ATP13A2.

(A) Representative western blots demonstrating pattern of Ub-proteins in SHSY5Y cell lines overexpressing WT, DN, sh-ATP13A2 and also an N-terminal fragment of ATP13A2 (N-ter long, first 251 amino acids). Following 24 h treatment with Bortezomib (10 nM) SHSY5Y cells were collected and the Ub pattern was assessed using immunoblotting of total cell lysate with an anti-Ub antibody. Actin served as the loading control. Data are representative of three independent experiments.

B) Ubiquitin intensities were normalized to actin, and plotted as fold change relative to Ctrl (untreated cells).

Figure 3. ATP13A2 dependent modulation of Ubiquitylated protein clearance under proteotoxicity, is not reliant on autophagy.

A) SYSH5Y cells were treated with Bortezomib (Borte, lOnM) for 24 h and cell lysates were collected for western blotting against LC3 to assess autophagy. Immunoblots are representatives of LC3-II formation in WT ATP13A2, catalytic inactive mutant (D508N), overexpressed the N-terminal part (AA 1 to 178) of ATP13A2 and sh-ATP13A2, in comparison to control. Although WT ATP13A2 overexpression triggers higher LC3-II levels in bortezomib conditions, this is not observed for the N-terminal part (AA 1 to 178), suggesting that the removal of Ub proteins by the N-terminal part (AA 1 to 178) occurs independently of autophagy. Data are representative of at least three independent experiments.

Figure 4. Mutations in the key lipid binding sites of ATP13A2 attenuate the ability of ATP13A2 to regulate Ub-protein clearance.

Representative western blots representing the ubiquitin accumulation pattern of SHSY5Y lines stably overexpressing WT ATP13A2, DN ATP13A2, LBS123 ATP13A2 (carrying mutations in the three N-terminal lipid binding sites LBS 1, 2 and 3 ) [Martin, S., et al., Parkinsons Dis, 2016. p. 9531917] versus control cell and sh-ATP13A2. Following a 24 h of treatment with Bortezomib (10 nM) in SYSH5Y, levels of Ubiquitylation were analyzed by immunoblotting with a primary anti-Ubiquitin antibody in comparison to actin as a loading control. Data are representative of at least three independent experiments.

Figure 5. Following Bortezomib (Borte) treatment, ATP13A2 overexpression enhanced secretion of nanovesicles in a lipid-dependent manner, which is recapitulated by the overexpression of a catalytic dead mutant or The N-terminal part (AA 1 to 178) of ATP13A2.

A) Quantification of exosome release in the media from SHSY5Y cells stably overexpressing ATP13A2 (WT, DN, LBS123) compared with sh-ATP13A2. Upon treatment with Bortezomib ( 10 nM, 24 h). Nano-vesicle fraction was purified via three different stages of centrifugation of cultured media at 4°C; 15 min at 500xg, 30 min at 10,000xg, 140,000xg for 3 h. Exosomes concentration was quantified and data were analysed by one-way ANOVA. N=3, *p<0.05

B) In conditions of Bortezomib ( 10 nM, 24 h), overexpression of The N-terminal part (AA 1 to 178) leads to an increased secretion of exosomes, whereas sh-ATP13A2 reduces the release of exosomes. Inhibition of PI(3,5)P2 formation with a PIKfyve inhibitor impairs the stimulatory effect of WT or N-ter, but had no effect on sh- ATP13A2.

C) Immunoblot analysis of exosome marker proteins (TsglO l, and CD9) and ATP13A2 release in exosomes from sh-ATP13A2 and ATP13A2-overexpressed SHSY5Y cells upon Bortezomib treatment ( 10 nM, 24h). Results were confirmed by two additional experiments.

Figure 6. Overexpression of ATP13A2 WT, DN, sh-ATP13A2 or The N-terminal part (AA 1 to 178) does not change the overall protein expression levels of key endo-/lysosomal modulators.

(A) Immunoblot representing SHSY5Y cell models stably overexpressing WT, DN, N-ter long (the N-terminal part (AA 1 to 178) of ATP13A2 versus) control and sh-ATP13A2 upon Bortezomib treatment. Expression levels of Rab5, Rab7, LAMPl were analyzed by immunoblotting with primary anti-Rab5, Rab7, LAMPl antibodies in comparison to beta-actin as a loading control. Data are representing at least three independent experiments.

(B,C) Bar graph representing the average ± SEM of the data in A. ^Statistical difference between non-treated versus Borte-treated in SHSY5Y (B,C), respectively (*p < 0.05; ANOVA with Dunnetts) (n = 3).

Figure 7. ATP13A2 overexpression in conditions of Bortezomib treatment results in a re-localization of a pool of LAMPl endo-/lysosomes out of the perinuclear region closer to the plasma membrane in line with the stimulation of exosome release via secretion A-C) Confocal images of the LAMPl immunostaining (as endo-/lysosomal marker) of cells overexpressing WT ATP13A2 (A) in comparison with sh-ATP13A2 silenced (B) SHSY5Y cell models Upon Bortezomib treatment ( 10 nM, 8h). Larger magnification of boxed areas (n = 10)

Figure 8. A schematic drawing of the ATP13A2 N-terminal part with indication of key elements in the sequence (P, putative phosphorylation sites; Ub, putative ubiquination sites; LBS, lipid binding sites).

Figure 9. ATP13A2 mediates protection against a spectrum of PD related agents. SHSY5Y neuroblastoma cells stably overexpressing ATP13A2 or shATP13A2 in comparison to Flue control were exposed to a dose response of the mitochondrial toxins; rotenone (A, E), MPP+ (B, F), 6-OHDA (C, G) and Rhodamine 6G (D, H) or the metal ions Zinc (I, M), Manganese (J, N), Iron (K, O) and Nickel (L, P) for 24 h. Treated cells were assessed for the effect of ATP13A2 expression on cell viability (A, B, C, D, I, J, K, L) and cell death (E, F, G, H, M, N, O, P). Data are the mean of a minimum of three independent experiments ± SD.

Figure 10. ATP13A2 mediated protection is dependent on lysosomal functionality. SHSY5Y based cell models of ATP13A2 expression were exposed to rotenone (1 μΜ) or zinc ( 100 μΜ) alone or in combination with Bafilomycin Al to inhibit lysosomes (Baf, 50 nM) for 24 h and assessed for cell death (A) and viability (B) . The data show that the loss of lysosomal functionality prevents the protective effect of ATP13A2 in conditions of Zinc or rotenone.

(C) Rotenone treatment incites ATP13A2 activity, which was assessed in ATP13A2 OE cells by assessing the phospho-intermediate formation (activity) of ATP13A2 in both the mitochondrial/lysosomal (M/L) and microsomal (μ) membrane fractions of SHSY5Y cells following 24 h exposure to rotenone ( 1 μΜ) . To verify the loading, the levels of ATP13A2 were compared via western blotting with an ATP13A2 specific antibody.

Figure 11. ATP13A2 protects against a spectrum of reactive oxygen species (ROS) inducing agents. SHSY5Y cells were exposed for 24 h to a spectrum of disease and mitochondrial associated agents (rotenone, MPP+, 6-OHDA, Zn2+, Mn2+, Fe3+ and Ni3+), which all lead to the generation of both peroxide (A, DCFDA) and superoxide (B, mitoSOX) based free radicals (assessed by flow cytometry) . C-F. To delineate the role of ATP13A2 in the cyto-protective response to ROS, SHSY5Y cells overexpressing WT-ATP13A2 were compared to Flue control and sh-ATP13A2 upon exposure to rotenone (C & D, 1 μΜ) or Zn ²⁺ (E & F, 150 μΜ) . ROS was followed over time for both DCFDA (C & E) and mitoSOX (D & F) by flow cytometry. Overexpression of ATP13A2 reduces the accumulation of ROS over time, whereas knockdown has the opposite effect.

Figure 12. ATP13A2 protection is activated by ROS and the protective effect of ATP13A2 can be mimicked by the addition of an anti-oxidant. To assess the role of ROS in ATP13A2 based cyto-protection; flue, ATP13A2 overexpressing and shATP13A2 SHSY5Y cell lines were exposed to rotenone (rot, 1 μΜ), Zinc (Zn, 150 μΜ), Manganese (Mn, 2 μΜ) or Iron (Fe, 1,5 mM) for 24 h, alone or in combination with mitoTEMPO (1 μΜ), and DCFDA (A) and mitoSox (B) were detected by flow cytometry. Furthermore, to correlate ROS to cellular stress and death, MUH based cell viability (C) and propidium iodide (D) assays were performed. Finally, to delineate the impact of ROS on mitochondrial functionality, western blot analysis was performed for key markers of the mitochondrial unfolded protein response (mtUPR), CHOP and HSP60 in comparison to GAPDH, showing that ATP13A2 overexpression reduces, whereas knockdown increases mtUPR (E) .

Figure 13. ATP13A2 minimizes rotenone induced mtUPR. We followed mtUPR induction as a marker for mitochondrial dysfunction and protein stress in the mitochondria . SHSY5Y cells with stable overexpression of (WT-OE) or reduction in (shATP13A2) ATP13A2, compared to Flue control were assessed over time for the effects of rotenone on mtUPR induction. Cells were exposed to 1 μΜ rotenone (rot) and harvested over time (0-24 h) and western blot analysis performed for the mitochondrial chaperone HSP60 and the transcription factor CHOP (A). Moreover, transcriptional activation of the mtUPR was confirmed by western blot analysis for co- induction of CEBP/Beta and CHOP at basal 8 h and 24 h post rotenone exposure (B). Figure 14. ATP13A2 mutations recapitulate mitochondrial phenotype in patient derived fibroblasts. Basal mitochondrial functionality/health of two independent patient derived fibroblast models of ATP13A2, possessing either the T512I or Phe851Cysfs*6 (DeITT) variants were characterized in comparison to wild-type controls (WT001, WT002) . Models were assessed for mitochondrial potential by TMRM based flow cytometry (A), ATP production using the commercially available luciferase based assay (B), ROS (C, mitoSOX ) and mitochondrial mass (D, MitoTracker Deep RED) based flow cytometry. mtUPR induction was assessed by western blot analysis of HSP60, CHOP and CEBP/Beta in comparison to BiP as a control of UPR and GAPDH as loading control (E).

Figure 15. ATP13A2 patient derived fibroblasts display heightened ROS dependent sensitivity to rotenone induced mitochondrial stress. Two patient derived fibroblast models of ATP13A2, possessing either the T512I or Phe851Cysfs*6 (DeITT) variants, demonstrated increased mitochondrial sensitivity to rotenone ( 1 μΜ), in comparison to wild-type controls, characterized by mass (A), potential (B) and ROS production (C) readouts. Moreover, ATP13A2 disease variants were more susceptible to rotenone ( 1 μΜ) induced cell death (D) which could be inhibited by pre-incubation ( 1 h) with the mitochondrial specific anti-oxidant mitoTEMPO (E, 1 μΜ). The capacity of mitoTEMPO to prevent rotenone induced cell death correlated directly with ROS scavenging (F). Figure 16. ATP13A2 promotes mitochondrial homeostasis independent of autophagy. To understand the modality by which ATP13A2 promotes lysosomal dependent mitochondrial homeostasis, SHSY5Y cells overexpressing ATP13A2 in comparison to flue control and shATP13A2 cells were exposed to rotenone (1 μΜ) and autophagy was assessed by western blot analysis for LC3 and p62 in comparison to actin as a loading control (A). As a positive control cells were treated with the mitophagy inducer CCCP ( 10 μΜ) or rotenone and autophagy was assessed (B). Moreover, ATP13A2 overexpressing and knockdown cells were transiently transfected with GFP-RFP-LC3 and the capacity of rotenone to incite autophagy assessed by confocal microscopy (C). Finally, ATP13A2 SHSY5Y cell models were exposed to CCCP and cell viability (E) and cell death were captured by MUH assay and Pi-based flow cytometry respectively. Figure 17. ATP13A2 promotes mitochondrial-lysosomal co-localization in a lysosomal functionality dependent manner. SHSY5Y cells with stable overexpression of Flue control (Α,Β), ATP13A2 overexpression (C, D; WT-OE) or knockdown (E, F; shATP13A2) were treated with rotenone (1 μΜ) alone or in combination with bafilomycin Al (Baf Al, 10 nM) and stained for lysosomes (LAMPl, GREEN) and mitochondria (TOMM20, RED) and co-localization (YELLOW, right section) was assessed by confocal microscopy. Co-localization coefficients (B, D, F) demonstrate increased co-localization in cells with increased ATP13A2.

Figure 18. ATP13A2 promotes mitochondrial-late endo-/lysosomal contact sites. To investigate the role of ATP13A2 in mitochondrial-lysosomal co-localization and potentiation of mitochondrial functionality, mitochondria were isolated from basal and rotenone (rot, 1 μΜ) treated SHSY5Y cells stably overexpressing Flue (control) in comparison to N-terminal (A) or ATP13A2 overexpressing (B) cell lines and western blot analysis was performed for FLAG or ATP13A2, TOMM20, and LAMPl. Treatment with rotenone promotes the presence of ATP13A2 in the mitochondrial fraction in line with contacts that are formed between lysosomes and mitochondria.

Figure 19. Purified N-terminal GST-labelled protein fragments of ATP13A2 LBS mutants stably interact with ATP13A2 which leads to inactivation of ATP13A2, whereas N-terminal GST labelled WT peptides stimulate activity of ATP13A2. The stable interaction might explain the dominant negative character of LBS1-3 mutants on ATP13A2 activity [Martin et al. cited above]. Microsomes harvested from SHSY5Y- hATP13A2 cells were incubated with purified N-terminal GST-labelled protein fragments of WT ATP13A2 and LBS mutants. We evaluated interaction of the peptides with full- length ATP13A2 via western blotting (A) and the effect thereof on ATP13A2 activity via an EP assay (B,C) . In (D), the much more stable formation of dimers is depicted for the LBS mutants as compared to WT. Note that LBS mutants are unable to autophosphorylate suggesting that the dimer is the inactive state. [Martin et al. cited above]

Figure 20. A schematic representation of the ATP13A2 N-terminus with highlighted epitope of a home-made, polyclonal ATP13A2 antibody (Ab; SY3072), membrane association region (Ma) and lipid binding sites (LBS). The Ab binds between LBSl/2 and LBS3, but closer to LBS3, the binding site for PA, which plays a strong role in ATP13A2 activation.

Auto-phosphorylation assay (EP) on microsomes, harvested from SHSY5Y-ATP13A2 WT cells, incubated with various dilutions of N-term ATP13A2 Ab (B), or two negative controls: a C-term ATP13A2 Ab (C) and an independent SPCAla Ab (D). EP assay was performed as described previously (Holemans et al, PNAS 2015). Only the N-term ATP13A2 Ab was able to increase phospho-enzyme formation of ATP13A2, presumably by preventing the binding of the N-terminus to the auto-inhibitory binding site.

Figure 21. A) The antibody SY3072 is directed against an epitope in the N-terminus of ATP13A2 ( 119-137) . B) HeLa cells are transiently cotransfected with hATP13A2 wild- type and the lysosomal/late endosomal marker RAB7. Immunocytochemistry was performed using ATP13A2 antibody SY3072 (dilution 1 : 250) and the Alexa fluor dye 546 ( 1 :2000). C) Mitochondrial-lysosomal fractions of COS cells transiently transfected with ATP13A2 or ATP13A2 fusion proteins (C-terminal V5 or GFP tag) were applied. As a negative control, the mitochondrial-lysosomal fraction of non-transfected COS cells was loaded. Blot was incubated with ATP13A2 SY3072 antibody ( 1 :2500) (left). Blots were incubated with ATP13A2 SY3072 antibody in presence or absence of the immunizing peptide (right).

Figure 22. ATP13A2 promotes a-syn clearance via exosomes. In a SHSY5Y cell model for PD with overexpression of alpha-synuclein, we now show that not only WT, but also disease mutants or catalytic dead mutants are equally potent in removing alpha- synuclein via exosomes.

Figure 23. shows a proof of concept for an assay to screen for ATP13A2 activating compounds.

A: The activity of P-type ATPases can be assessed by measuring ATP hydrolysis. Therefore, we use the ATPase assay. Samples are prepared with purified protein and ATP and left for 20-30 min for ATP hydrolysis to occur. Mo is added which will form a complex with Pi released during ATP hydrolysis. This complex is reduced by ascorbic acid to form the blue coloured β-keggin ion. Arsenite is added to complex the remaining molybdate. The amount of the blue coloured ion produced is proportional to the amount of Pi present and the absorption can be measured using a colorimeter to determine the amount of phosphorus. We will screen for activating compounds by following an increase in ATPase activity. Note that other detection methods to measure ADP or Pi production can be deployed. Alternatively, one can use a luciferase based system to monitor the ADP production, which is converted by ATP by an enzymatic reaction, which then stimulates luciferase (light emission) activity.

B: From a small screening, two compounds, vVl and vV2, are both able to induce ATPase activity of ATP13A2. In combination with the N-terminal interacting lipid PA, ATPase activity is stimulated even further. These compounds are used as positive controls in an ATPase assay suitable for high throughput screening. The activation by PA further supports our model of the N-terminal binding of the lipid that is required to unlock ATP13A2 to promote activity. Note that PA on itself is insufficient to promote ATPase activity. It depends on a substrate for transport.

Figure 24 Phospholipids PA and PI(3,5)P2 stimulate ATPase activity of purified ATP13A2.

A. The ATPase activity of purified recombinant ATP13A2 from yeast in the presence of a human metabolite is stimulated by addition of PA or PI(3,5)P2. Combining PA and PI(3,5)P2 in the assay leads to the highest ATPase activity.

B. The ATPase activity of ATP13A2 in microsome fractions of SHSY5Y neuroblastoma cells overexpressing ATP13A2 is stimulated by addition of PA to the membrane fractions.

These experiments provide a proof of concept that targeting the N-terminal auto- inhibitory domain of ATP13A2 leads to activation of activity. This assay can be used to screen for modulators of the ATP13A2 auto-inhibitory mechanism. The ATPase activity was measured with a luciferase-based protocol to measure the produced ADP concentrations over time.

In some embodiments, the fragment of ATP13A2 comprises a polypeptide sequence consisting of a sequence selected from :

SEQ ID Sequence Description

NO

1 Msadssplvg stptgygtlt igtsidplss vssvrlsgy N terminus with lipid cgspwrvigy hvvvwmmagi plllfrwkpl gvrlrlrpc interaction sites N- nlahaetlvi eirdkedssw qlftvqvqte igegsleps terminus with lipid pqsqaedgrs qaavgavpeg awkdtaqlhk eeavsvgqk interaction sites: residue rvlryylfqg qryiwietqq afyqvslldh gr 1- 187 (variant 2 or 3), residue 1-192 (variant 1)

2 Msadssplvg stptgygtlt igtsidplss Complete N-terminus until svssvrlsgy cgspwrvigy hvvvwmmagi SEQ ID Sequence Description

NO

plllfrwkpl wgvrlrlrpc nlahaetlvi M l : 1-225 (variant 2 or 3), eirdkedssw qlftvqvqte aigegsleps

residue 1-230 (variant 1) pqsqaedgrs qaavgavpeg awkdtaqlhk

seeavsvgqk rvlryylfqg qryiwietqq

afyqvslldh grscddvhrs rhglslqdqm

vrkaiygpnv isipvksypq

3 msadssplvg stptgygtlt igtsidplss Complete N-terminus also svssvrlsgy cgspwrvigy hvvvwmmagi

including M l : residue 1- plllfrwkpl wgvrlrlrpc nlahaetlvi

eirdkedssw qlftvqvqte aigegsleps 256 (variant 1), residue 1- pqsqaedgrs qaavgavpeg awkdtaqlhk 251 (variant 2 or 3) seeavsvgqk rvlryylfq gqryiwietqq

afyqvslldh grscddvhrs rhglslqdqm

vrkaiygpnv isipvksypq llvdealnpy

ygfqafsial wladhy

Fragments may be derived from human ATPase isoforms, such as

- isoform 1 (NP_071372.1) and splice variants (hATP13Al - splice-variant 1 :

Q9HD20.2; hATP13Al - splice-variant 2 EAW84854.1)

- isoform 2 (NP_001135446.1),

- isoform 3 (NP_001135446.1) and splice variants (hATP13A3 splice-variant 1 XP_011511423.1; hATP13A3 splice-variant 2: XP_011511424.1h; ATP 13 A3 splice-variant 3 XP_011511425.1; hATP13A3 splice-variant 4 XP_011511426.1) isoform hATP13A4 and splice variants (hATP13A4 splice-variant 1 NP_115655.2; hATP13A4 splice-variant 2 XP_016862807.1; hATP13A4 splice- variant 3 XP_011511534.1; hATP13A4 splice-variant 4 XP_016862808.1), isoform hATP13A5 and splice variants (hATP13A5 splice-variant 1 NP_940907.2; hATP13A5 - splice-variant 2 XP_011511072.1)

Sequences are from the NCBI Reference Sequence database, as updated on June 26, 2016. Fragments may also be derived from other orthologs of ATP13A2, such as ATP13A2 from primates, rodents, canines, or other mammals. EXAMPLES 1 Material and Methods

1.1 Cell Culture

SHSY5Y neuroblastoma, were cultured in DMEM (Dulbecco's Modified Eagle Medium) (Sigma-Aldrich, D6546) containing 1% Glutamine, 1% non-essential amino acids, 0.1% Gentamicin (Sigma-Aldrich, G7513, M7145, G1264, respectively) as well as 15% fetal bovine serum (FBS) (GE healthcare, Healthcare systems, Diegem, Belgium) for SHSY5Y. All cell lines were maintained in the incubator at 37° C and 5% CO2 for a maximum of 10 passages and regularly checked for mycoplasma contamination.

1.2 DNA construct of ATP13A2

SHSY5Y neuroblastoma cell lines stably expressing firefly luciferase (FLUC, control), WT ATP13A2, D508N ATP13A2 (DN), flag-labelled N-terminal long fragment of ATP13A2 (251 amino acids) and the LBS123 mutant with mutations in the three putative N-terminal Lipid Binding Sites (LBS123 :65FRWKP [SEQ ID NO:9] → FAWAP [SEQ ID NO: 10] ;74RLRLR [SEQ ID NO:4]→ ALALA [SEQ ID NO: 11] ; 155KRVLR [SEQ ID NO: 5] → AAV LA [SEQ ID NO: 12] ) or sh-ATP13A2 (KD) [Martin, S., et al., Parkinsons Dis, 2016. 2016: p. 9531917; Holemans, T., et al. Proc Natl Acad Sci U S A, 2015. 112(29) : p. 9040-5.] were generated via lentiviral transduction and maintained as described in Holemans et al., 2015.

1.3 Transient transfection with GFP-Ub and RFP-GFP-LC3.

ATP13A2 modulated cell lines were transiently transfected ( 1 μg : 3 μΙ; ratio of DNA/ lipofectamine) with either GFP-Ub or RFP-GFP-LC3 using lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA, 12566014) reagent and diluted in 100 μΙ FBS-free medium. Following 15 min incubation at room temperature, the diluted DNA was pipetted into the medium of seeded cells (lx 10 ⁵ cells per well). The medium of transfected cells were kept for few hours and then 1 ml full media was added on them.

1.4 Drug treatment

Drug treatments were performed in the medium, as explained above, without the addition of selective antibiotics. To inhibit the proteasome, cells were treated with Bortezomib (Borte, 10 nM) for a maximum of 24 hours. To observe the autophagic flux, cells were preteated for 1 h with Bafilomycin Al (ΙΟηΜ, InVivoGen, DMSO). -Cells were exposed to rotenone (Rot, 1 μΜ ; R8875, Sigma), MPP+ (50 μΜ, D048, Sigma), Zinc (ZnCI2, 150 μΜ; Z0152, Sigma), Manganese (MnCI2, 2 μΜ ; 205891000, Acros Organics) and Iron (FeCI3, 1.5 mM; 157740, Sigma) for 24 hours (h). The final concentration of each agent was chosen from a dose response analysis to obtain a sub- maximal inhibition of cell viability. Prior to stressor addition, cells were pre-treated for 1 h with YM-201636 (PIKfyve inhibitor, PIK, 200 nM; 524611, Millipore) or 5-Fluoro-2- indolyl des-chlorohalopemide (PLD inhibitor, FIPI, 100 nM; F5807, Sigma) to inhibit the production of PI(3,5)P2 or PA, respectively.

1.5 Preparation of cell extracts

lx 10 ⁶ SHSY5Y cells were seeded per 6 cm culture dish. Following treatments, cells were harvested via physical scraping, and cell pellets were washed twice with phosphate-buffered saline (PBS, Sigma-Aldrich, Bornem, Belgium) and lysed in RIPA buffer (Enzo life sciences, 80-1284) (50-75 μΙ depending on pellet volume) supplemented with protease and phosphatase inhibitors (Thermo Scientific, Waltham, MA, USA). Cell lysates were incubated on ice for 30 minutes and subsequently centrifuged at 18,000xg, 30 min, 4°C. Protein concentrations were determined using a commercially available BCA kit (Thermo Scientific, Waltham, MA, USA, 23228). 1.6 Western blot

To assess protein expression and induction, 50 μg of protein were separated using SDS-PAGE. Afterwards, proteins were transferred onto nitrocellulose Membranes (GE healthcare, Healthcare systems, Diegem, Belgium) and subsequently blocked with 5% non-fat milk in TBS-Tween ( 10 mM Tris, pH 8.0, 150 mM NaCI, 0.5% Tween 20) for 1 h prior to the addition of an array of primary antibodies and incubated overnight 4°C.

Primary antibodies were labelled with fluorescent secondary antibodies, including anti- rabbit-DyLight 800 and anti-mouse-Daylight 680 (Thermo Scientific, Waltham, MA, USA, 35571 and 35519) at a 1 : 2000 dilution. Images were taken and quantified using an Odyssey infrared imaging device (Li-Cor Biosciences, Lincoln, NE, USA) and analysed by ImageJ 1.49v and Graph Pad Prism 6.

1.7 Immunostaining

ATP13A2 cell models were plated at a density of lxlO ⁵ cells per well and cultured on cover slides in 12 well-plates. Cells were then treated overnight. At the relevant time points, cells were fixed with 4% paraformaldehyde (Sigma-Aldrich, Bornem, Belgium) at 37° C for 20 mins. Cellular permeabilization was induced with 0.1% Triton X-100 (Sigma-Aldrich, Bornem, Belgium) for 10 min at room temperature. Afterwards, coverslips were blocked in PBS containing 1% BSA (Sigma-Aldrich, Bornem, Belgium) and 10% FBS for 1 h at room temperature, followed by the incubation with primary antibodies; ATP13A2 and LAMP1 or CD63, diluted in blocking solution at 4° C overnight. Samples were then washed twice with PBS, and stained coverslips were labelled with Alexa Flour 488- or Alexa Flour 546-conjugated secondary antibodies for 1 h in RT. Nuclei were also counterstained with DAPI ( 1 μg/MI) (Sigma-Aldrich, Bornem, Belgium), mounted using Prolong Gold anti-fade reagent and cured overnight. Confocal images were captured using confocal microscopy ZEISS Axiovert 100M LM510 and Zeiss plan-Apochromat 63x : 1.4 NA oil immersion objective and/or Nikon AIR Eclipse Ti with Apo 60X ( 1,40 oil) objective. 1.8 Cellular fractionation

Cells were seeded in 15 cm dishes at a density of 1.5 x 106. After treatment, cells were harvested and resuspended in hypotonic buffer [ 10 mM Tris-HCI pH 7.5, 0.5 mM MgCI2, SigmaFast protease inhibitor cocktail (Sigma)] and a 10 min incubation period on ice. Cells were homogenized by applying 40 strokes in a Dounce homogenizer and after adding [0.5 M sucrose, 10 mM Tris.HCI pH 7.3, 40 μΜ CaCI2, 0.23 μΜ phenylmethylsulfonyl fluoride, 1 mM dithiothreitol (DTT)] 20 additional strokes were performed. The total cell lysate was subjected to differential centrifugation : the nuclear fraction (1,000 g; 10 min), the mitochondrial/lysosomal fraction ( 12,000 g; 20 min) and lastly, the microsomal and cytosolic fractions (200,000 g; 35 min; pellet and supernatant, respectively). Microsomal pellets were resuspended in a 250 mM sucrose solution supplemented with protease inhibitor cocktail. All fractionation steps were carried out at 4°C. After solubilization, samples were aliquoted, flash-frozen in liquid nitrogen and stored at -80°C. The protein concentration was determined by the Qubit fluorimetric method (Life Technologies). 1.9 Nano-vesicle isolation

To isolate Nano-vesicles, SHSY5Y cell lines were cultured with initial equivalent cell number (4x l0 ⁶ / plate). Upon reaching semi-confluency, cells were washed two times with PBS and refreshed with DMEM consisting of 1% exosome depleted FBS with defined treatments. 24 h later, the cell-conditioned media was collected from triplicate cultures. Nano-vesicles were purified using a differential centrifugation method at 4°C: 15 min at 500xg to remove cells; 30 min at 10,000 to eliminate cell debris; filtration of apoptotic bodies with 0.22 μιη filter (Merck Millipore, Darmstadt, Germany) and final centrifugation at 140,000xg for 3 h to pellet nano-vesicles. Supernatant was decanted and nano-vesicles collected in 50 μΙ RIPA buffer.

1.10 Autophosphorylation assay

40 μg of the microsomal fraction was added to a final volume of 95 μΙ of reaction buffer (160 mM KCI, 17 mM Hepes, 2 mM MgCI2, 1 mM DTT, 5 mM NaN3). The autophosphorylation assay was started by adding [y-32P] ATP (2 μθ) and stopped after 1 min with 400 μΙ of ice-cold stop solution (20% trichloroacetic acid, 10 mM phosphoric acid). Samples were incubated on ice for 30 min to precipitate protein and centrifuged at 20,000 g for 30 min at 4°C. The pellet was washed twice with 400 μΙ of ice-cold stop solution and finally, dissolved in sample buffer ( 10% LDS, 10 mM NaH2P04, 0.01% SDS, 10 mM 2-mercaptoethanol, 0.15 mg/ml bromophenol blue). After loading the samples on NuPAGE 4-12% BisTris gels (Life Sciences), electrophoresis was conducted for 1.5 h (40 mA, 170 V) in running buffer containing 0.1% SDS and 170 mM MOPS (pH 6.3). Following fixation in 7.5% acetic acid, the gel was exposed to a Phosphorlmager screen (GE Healthcare) and the next day, radioactivity was visualized in a Phosphorlmager scanner (Storm 860, GE Healthcare). Quantification was performed with Image QuanT (Molecular Dynamics) and ImageJ software packages (http://rsbweb.nih.gov/ij/).

1.11 Cell death assay Cell death was determined by propidium iodide (PI) exclusion. Briefly, cells were trypsinized at the indicated time points and incubated with PI. Pi-positive (dead) cells were quantified via flow cytometry (Attune Cytometer, Life Technologies).

1.12 Cell viability assay

Cells were seeded at 5000 cells per well of a 96-well plate. Following treatment cells were washed with PBS and incubated with 0.01 mg/ml MUH (4-methylumbelliferyl heptanoate, Sigma) dissolved in PBS for 30 min at 37 °C. Fluorescence was measured with a Flex Station plate reader (Molecular Devices) with excitation 355 nm, emission 460 nm and cut off value of 455 nm.

1.13 Statistical analysis

Data were presented at the average ± SEM, of at least three independent experiments. Statistical analysis was conducted using Graph Pad Prism 6 by ANOVA with a Bonferroni post hoc test. Considered significance * ^/$ p < 0.05, "^p < 0.01, and *** ^/ **i < 0.001. Computer package of ImageJ 1.49v was used for densitometry analysis of immunoblot signals.

2 Results

2.1 ATP13A2 is primarily expressed and localized in the late endosomes/ lysosomes.

ATP13A2 expression levels were investigated for SHSY5Y cell models stably overexpressing wild type (WT) and catalytically inactive D508N (DN) variants of ATP13A2, in comparison to sh-ATP13A2 knockdown and firefly-luciferase overexpressing control cell lines (Figure. 1) . ATP13A2 WT and DN were shown to be significantly overexpressed (Figure. 1A). Moreover, cellular localization of all cell lines was examined by co-localization studies, using LAMP2 as a positive marker of the late endosomes/lysosomes. Imaging showed that ATP13A2 co-localized with LAMP2 positive vesicles indicating that ATP13A2 is expressed and localized in the late endosome/lysosomes compartments. Knocking down of ATP13A2 in our cell line models is confirmed by measuring mRNA levels with Q-RT PCR (mRNA levels were reduced by 75.4 ± 8.9%) and via mass-spectrometry at the protein level.

2.2 Upon loss of proteostasis, ATP13A2 mediates clearance of ubiquitinated proteins through mechanisms that are unaffiliated to its catalytic activity. To decipher the cellular implication of ATP13A2 in protein quality control (PQC) processes under conditions of proteotoxic stress, SHSY5Y cellular models were treated with the proteasome inhibitor Bortezomib and the pattern of total intracellular ubiquitin accumulation was examined via western blot analysis and GFP-labelled Ubiquitin immunofluorescence (Figure. 2). Interestingly, the up-regulation of ATP13A2 expression in SHSY5Y alleviated the intracellular accumulation of ubiquitinated proteins (Figure. 2A). To implicate the catalytic activity of ATP13A2 in Ub-based clearance we investigated the clearance capacity of the ATP13A2 catalytically dead mutant, with a key mutation in the aspartate residue of the enzyme catalytic site (D508N). Interestingly, Ub-proteins clearance happened independent of ATP13A2s catalytic activity, as there was a comparable Ub-phenotype both in WT-OE cells and DN-OE cells (Figure. 2A-B). A study published by Holemans et al., uncovered a notable role of the unique N- terminus domain of ATP13A2, where it acts as a key regulator of ATP13A2's activity. This regulation was found to be mediated by ATP13A2's direct interaction with phosphatidic acid (PA) and phosphatidylinositol (3,5) biophosphate (PI(3,5)P2), two stress inducible signalling lipids involved in the endo-/lysosomal pathway, trafficking, acidification and autophagy. Moreover, these interactions occurred through the N- terminus domain. Interestingly, we found that cells overexpressing only the N-terminal domain of ATP13A2 phenocopied the Ub-proteins clearance phenotype observed for WT/ DN OE cell lines (Figure. 2A) . Thus, our observations suggest a role for ATP13A2 in the Ub-protein turnover/clearance in cells responding to Bortezomib, which is exerted in a catalytically independent, but N-terminus dependent manner. Although catalytically inactive, the N- terminus may act as a scaffold protein recruiting a protein complex mediating ubiquitinated protein clearance via the endosome/lysosome. 2.3 Regulation of ubiquitylated-proteins clearance by ATP13A2 following loss of proteostasis is not reliant on autophagy

As a prominent role of autophagy-mediated degradation is the maintenance of cell homeostasis and as ATP13A2 potentially localizes to the outer limiting membrane of MVB we evaluated the contribution of autophagy in the ATP13A2-mediated clearance of Ub-proteins (Figure. 3). To evaluate autophagic flux alterations, firstly, we monitored LC3 (Microtubule-associated protein light chain 3) conversion upon treatment with Bortezomib. As previously reported, Bortezomib treatment increased LC3-II formation in all cell models however this was significantly higher in OE cells, both for the WT and DN cell lines (Figure. 3A), opposite of the Ub clearance patterns (Figure. 2). To confirm these observations these cell lines were transfected with GFP- LC3 and the formation of LC3 puncta were monitored via immunostaining. Elevated GFP-LC3 granularity was evident in the cell lines overexpressing ATP13A2.

Interestingly, however, flux analysis, measured by the addition of bafilomycin Al, did not reveal a further increase in LC3-II formation, suggesting that the observed accumulation of LC3-II was not a direct result of enhanced autophagic flux but conversely, because of lowered LC3-II lysosomal turnover/degradation. In line, we found an enhanced amount of uncleared autophagosomes in the ATP13A2 OE cells, correlating directly with our previous observations that ATP13A2 reduces autophagic flux. Finally, to correlate the Ub-clearance phenotype with autophagy, we investigated the capacity of the N-terminal of ATP13A2 to incite or regulate autophagy. Interestingly, overexpression of the N-terminal domain of ATP13A2 did not result in LC3-II accumulation, as observed for ATP13A2 OE cells, upon treatment with Bortezomib (Figure. 3A), in spite of the similar effects observed for the Ub-protein clearance phenotype.

Taken together, these data show that ATP13A2 regulates Ub-protein clearance through its N-terminus, which occurs independently of autophagy. However, ATP13A2 facilitates a cellular response that is detrimental to the overall removal of LC3-II positive vesicles.

2.4 ATP13A2-mediated Ub-proteins turnover is regulated by its scaffolding capability through the N-terminus interaction with lipids.

As the N-terminus of ATP13A2 interacts with the lysosomal signaling lipids PA and PI(3,5)P2 to regulate activity we investigated the role of these lipids in our Ub- clearance phenotype (Figure. 4). To that end, we modulated the levels of both PA and PI(3,5)P2 by the pre-treatment with YM-201636 (a PIKfyve kinase inhibitor, PI(3,5)P2) or FIPI (a PLD1 inhibitor, PA) and monitored the Ub-phenotype incited by Bortezomib . Our data showed that upon blockade of PI(3,5)P2, the accumulation of Ub-proteins accelerated in all of ATP13A2 overexpressing cells (WT OE, DN OE, N-ter long OE), yet had no significant effect on the sh-ATP13A2 cell lines. Moreover inhibition of PI(3,5)P2 formation alone had no influence on Ub-proteins turnover. Interestingly, FIPI induced reduction in PA had no significant effect on ATP13A2 mediated removal of Ub-proteins.

To verify the relevance of the lipid interactions in ATP13A2 Ub-proteins phenotype, we used stable cell lines overexpressing lipid binding site mutants in the SHSY5Y cell model overexpressing an ATP13A2 mutant, in which the three N-terminal lipid binding sites are mutated (LBS123 OE) . In line with the data obtained by pharmacological inhibition, blocking ATP13A2's capacity to interact with PA or PI(3,5)P2, decreased its ability to mediated Ub-protein clearance (Figure. 4) . These data indicate that ATP13A2 possesses significant lipid sensitive scaffolding potential that acts independent of its catalytic activity to modulate the removal of Ub- proteins during proteotoxic stress. Moreover, ATP13A2's scaffolding function is modulated by PI(3,5)P2.

2.5 ATP13A2 removes Ub-proteins by means of endocytic export in a PA/PI(3,5)P2- dependent manner

As a consequence of ATP13A2s previous links with MVBs and the increased secretion of alpha synuclein, coupled with our findings that autophagy was not the major route of ATP13A2-mediated Ub-clearance, we investigated the role of the endo-/lysosomal pathway (Figure. 5,6).

Therefore, to determine whether the overexpression of ATP13A2 enhances the externalization of Ub-proteins via nano-vesicles, we purified crude nano-vesicles fractions from SHSY5Y cell lines. Initial data demonstrated that, following Bortezomib exposure, the amount of secreted nano-vesicles proteins was enhanced in the WT-OE and DN-OE cell lines. Furthermore, exposure to Bortezomib had no significant effect on the secretion of nano-vesicles by either sh-ATP13A2 or LBS123 OE cell lines. Remarkably, pre-treatment with the PIKfyve inhibitor, prevented the release of nano- vesicles observed in both the WT-OE and DN-OE cell lines. Subsequent analysis of the released nano-vesicle fractions found them to be positive for both exosomal markers TSGlOl and CD9. Remarkably, ATP13A2 was also released strongly in the nano-vesicle fractions of WT-OE cells of SHSY5Y cell lines, only following treatment with Bortezomib (Figure. 5B) .

To further delineate the potential of ATP13A2 to modulate vesicular secretion of Ub- proteins, we defined intracellular protein expression levels of the key endo-/lysosomal modulators Rab5, Rab7 and Lampl (Rab5; the early endosomal marker, Rab7; the late endosomal marker, LAMP1; the lysosmal marker) . Neither ATP13A2 expression nor Bortezomib exposure affected the overall expression level of these endo-/lysosomal markers (Figure. 6A-C).

2.6 ATP13A2 overexpression results in a re-localization of a pool of endo- /lysosomes out of perinuclear region

Since ATP13A2 had no effect on the expression of the endo/lysosomal machinery, we hypothesized that ATP13A2 may modulate the movement of the endocytic compartments to the plasma membrane, promoting the release of Ub-proteins during periods of proteotoxicity. Co-localization studies in SHSY5Y cell lines (WT OE, LBS123 OE versus sh-ATP13A2 cell) showed that, following 8 h incubation with Bortezomib, a significant re-distribution of the endocytic network occurred. More specifically, ATP13A2 OE modulated the trafficking of CD63/LAMP1 positive vesicles out of the perinuclear region in the cell (Figure 7 A-B). Taken together, these findings strongly indicate that, in cells subjected to proteotoxic stress, ATP13A2s scaffolding function leads to the decreased accumulation of intracellular Ub-proteins by redistributing a portion of CD63/LAMP1 positive vesicles out of the perinuclear region, correlating with the detection of increased nano-vesicle secretion. 3 Discussion

The observations derived from this study highlight a novel role for the P5-type ATPase ATP13A2 in endo-lysosomal mediated removal of ubiquitinated proteins that occurs independent of the catalytic turnover of ATP13A2. Moreover, this is a direct consequence of endocytic remodelling towards a secretory phenotype, for which ATP13A2s novel scaffold function and the interaction with the endocytic signalling lipid PI(3, 5)P2 is required.

3.1 ATP13A2, a dynamic modulator of protein quality

Initially our study showed that ATP13A2 co-localized to LAMPl-positive vesicles, providing further evidence that ATP13A2 resides within the late endo/lysosomal system. Previously, co-localization studies showed ATP13A2 co-localization with Rab5 (a marker of early endosomes), Rab7 (a marker of late endosomes), LC3 (a marker of autophagosomes) [Ramonet, D., et al., Hum Mol Genet, 2012. 21(8) : p. 1725-43.; Kong, S.M ., et al., Hum Mol Genet, 2014. 23(11) : p. 2816-33.] CD63 (a marker of LE/MVBs as well as exosomes) [Tsunemi, T., K. et al. J Neurosci, 2014. 34(46) : p. 15281-7] and lastly, MAP2 and βΙΙΙ tubulin (markers of the microtubule network) [Ramonet cited above]. Taken together, these observations suggest that ATP13A2 may reside on the membrane of intracellular vesicular compartments, including late endosomes, lysosomes and exosomes [Ramonet cited above]. Autophagosomes can interact with late endosome and lysosomes to mediate the removal of cellular components and organelles, however Kong et al, demonstrated that ATP13A2 co- localization with LC3 occurred independent of lysosomal functionality, as following exposure to chloroquine ATP13A2 and LC3 co-localization remained only partial [Kong, S.M., et al., Hum Mol Genet, 2014. 23(11) : p. 2816-33]. These observations occurred as a consequence of autophagosome fusion with the late endosomal compartment better known as the amphisome [Kong et al cited above] . This highlights the potential role of ATP13A2 as a central player within cellular proteostasis mechanisms. 3.2 ATP13A2, a key player in Ub-protein homeostasis

The observations that, upon proteostasis challenge in SHYS5Y, ATP13A2 OE cell lines accumulate less ubiquitinated proteins highlights that ATP13A2 has besides a transport function also a role as a scaffold. We observed that this phenotype was independent of activity as the catalytic inactive variant as well as the N-terminus of ATP13A2 reduced Ub-protein accumulation. Together these data support the idea that ATP13A2 is a key player in Ub-protein turnover while this role supports a scaffolding, not catalytic role, exerted through the N-terminus of ATP13A2.

ATP13A2 is known to protect against toxicity of heavy metals including zinc [Kong, S.M., et al., Hum Mol Genet, 2014. 23( 11) : p. 2816-33.; Tsunemi, T. and D. Krainc, Hum Mol Genet, 2014. 23(11) : p. 2791-801], iron [Rinaldi et al. (2015) Biochim Biophys Acta, 1848, 1646-55] manganese, nickel, zinc [Gitler cited above; Covy et al. 2012 J. Neurosci. res 90, 2306-2016] as well as oxidative stress and mitochondrial impairment [Covy et al cite before; Gusdon, A.M ., et al., Neurobiol Dis, 2012. 45(3) : p. 962-72] . Interestingly, research has linked ATP13A2 with the enhanced secretion of a-synuclein [Kong et al. cited above]. Within this study, we provide evidence that the secretory capacity of ATP13A2 may in fact be more general, and the removal of a- synuclein may be linked to ubiquitination. Unfortunately, the study that identified ATP13A2's secretory potential failed to consider an activity independent function. Therefore all these roles together, may describe a general protective effect of ATP13A2, specifically in neurons, in which ATP13A2 expression can prevent toxicity, for example by a-synuclein [Gitler et al. cited above, Usenovic et al. cited above] .

3.3 ATP13A2 activity via its N-terminus and PA and PI(3,5)P2 interaction

As mentioned, Holemans et al demonstrated that ATP13A2 protein has a unique hydrophobic N-terminus interacting with the stress inducible lipids PA and PI(3,5)P2, which are able to regulate the activity of ATP13A2 under cellular stress conditions [Holemans et al. cited above] . Moreover, this protective effect of ATP13A2 is essential under periods of mitochondrial stress as when the availability of PA and PI(3,5)P2 phospholipids was diminished, ATP13A2 could no longer protect against rotenone induced toxicity [Holemans et al. cited above]. Similarly, a recent study showed that the protective effect of ATP13A2 against metal toxicity declined when mutations in the PA/PI(3,5)P2 binding sites of ATP13A2 N-terminus were applied [Martin et al cited above]. Comparably, within this study we observed that the N-terminal domain of ATP13A2 alone was sufficient to phenocopy and incite the Ub-protein clearance phenotype observed for the full-length ATP13A2 variants.

Consistent with previous findings, mutation in the PA/PI(3,5)P2 binding sites impairs the ATP13A2-mediated phenotype of Ub-protein removal leading to an accelerated accumulation of Ub-proteins compared with the ATP13A2 WT-OE cell lines. Together, these observations highlight the critical importance of the lipid switch in ATP13A2 activation. This raises important questions into the role of these phospholipids within the intermolecular functionality of ATP13A2.

We propose a model of ATP13A2 dimerization, which represents an inactive resting state unable to exert its transport and scaffold functions. Upon a dimer to monomer transition that is controlled via the N-terminus by the interaction with the N-terminal lipids PA and/or PI(3,5)P2, ATP13A2 becomes activated, which provides protection to mitochondrial or metal-induced stress. In addition, in the monomeric state, the N- terminus is available to exert the scaffold function, which promotes the clearance of ubiquitinated proteins. By overexpression of an N-terminal fragment of ATP13A2, the dimer to monomer transition is by-passed, leading to a constitutive activation of the scaffold function. Thus promoting the dimer to monomer transition or expressing monomeric forms of ATP13A2 or ATP13A2 fragments might be potent strategies to improve the clearance of ubiquitinated proteins, such as alpha-synuclein. 3.4 The role of ATP13A2 in the exporting of Ub-protein

Interestingly, here we only see sensitivity to the loss of PI(3,5)P2. PI(3,5)P2 is an essential trafficking lipid, that in specialized scenarios is required for cellular trafficking along cytoskeletal filaments. For example, cortactin interacts directly with PI(3,5)P2 and mediates actin turnover at the level of the late endosomes [Hong, N .H., et al. J Cell Biol, 2015. 210(5) : p. 753-69] . Moreover, PI(3,5)P2 has essential roles in the regulation of lysosomal motility, due to the interaction with dynein motors and is necessary for lysosome tubulation and reformation [ Li, X., et ai. , Nat Cell Biol, 2016. 18(4) : p. 404- 17] .

Based on our findings, ATP13A2 mediated removal of ubiquitylated proteins occurred independent from autophagy. Consequently, we observed that upon proteotoxic stress, overexpression of ATP13A2 led to a shift in the localization of a pool of LEs/lysosomes, from the degradative (perinuclear) toward a fraction that may be more active in cargo movement and export. Interestingly, we observed that during proteotoxicity, the overexpression of ATP13A2 reduced autophagosomal-lysosomal fusion which may have been an artefact of endocytic remodelling, the movement of lysosomes away from the autophagosome and therefore giving the interpretation of a reduced cellular autophagic flux capacity.

The pathway by which ATP13A2 leads to this shift, is still unanswered. Due to the existence of ATP13A2 in different vesicles like EE, LE, lysosomes, MVBs, exosomes and even microtubule vesicles, various hypotheses may be raised around the involvement of ATP13A2 in Ub-protein clearance. However, it is already displayed that the loss of ATP13A2 in the outer membrane of MVBs results in the MVBs dysfunction and that ATP13A2 facilitates the secretion of a-synuclein [Tsunemi, T., K. Hamada, and D. Krainc, Neurosci, 2014. 34(46) : p. 15281-7.] Moreover, sorting of ubiquitylated cargo into the MVBs for export as exosomes, is also associated with the ESCRT complexes ability to bind to PI(3,5)R2 [McCartney, A. J., et al. Bioessays, 2014. 36( 1) : p. 52-64.]. On the other hand, Kong and co-workers observed that the fractionation patterns of increased exosome-associated a-synuclein was mirrored by HSP70 overexpression, the key chaperone in exosome cargo sorting [Kong et al cited above]. Considering the equivalent function for Hsp70 with Hsc70, a chaperone mediating EMA, elevated ATP13A2 expression might play a significant role in the regulation of EMA.

Moreover, the release of ATP13A2 in nano-vesicles of cells overexpressing ATP13A2, raise the possibility that ATP13A2 may play a role in regulation of exosomes biogenesis and their release.

3.5 The role of ATP13A2 in PD

We have demonstrated that enhanced ATP13A2 expression strives to decrease intracellular Ub-proteins and increment their externalization within nano-vesicles, such as exosomes, under conditions of proteasomal stress. While ATP13A2 knock down resulted in enhanced intracellular accumulation of Ub-proteins.

Released exosomes, which are picked up by neighbouring cells, are the potential means of cell-cell communication [Thery, C. et al. Nat Rev Immunol, 2002. 2(8) : p. 569-79. ; Urbanelli, L, et al., Signaling pathways in exosomes biogenesis, secretion and fate. Genes (Basel), 2013. 4(2) : p. 152-70.] . The broadening role of exosomes is being hunted throughout cellular pathology of different diseases, such as neurodegeneration. Recently studies in the field of neurodegeneration have demonstrated that on one hand, the export of aggregated toxic proteins, via exosomes can be beneficial to the dopaminergic neurons and result in postponed onset of PD. However this could in fact deliver toxicity to the neighbouring neurons and promote PD progression [Kong et al. cited above], amplified by a prion disease like transmission. However, the pathological role of exosomes in vivo is still not fully understood.

INCORPORATION BY REFERENCE

All publications and patents mentioned herein are hereby incorporated by reference in their entirety as if each individual publication or patent was specifically and individually indicated to be incorporated by reference. In case of conflict, the present application, including any definitions herein, will control.

EQUIVALENTS

While specific embodiments of the subject invention have been discussed, the above specification is illustrative and not restrictive. Many variations of the invention will become apparent to those skilled in the art upon review of this specification and the claims below. The full scope of the invention should be determined by reference to the claims, along with their full scope of equivalents, and the specification, along with such variations.