VCP/p97 mediates nuclear targeting of non-ER-imported prion protein to maintain proteostasis

Non-ER-imported PrP is targeted to the nucleus

To specifically study cytosolic quality control pathways that operate on the fraction of secretory proteins that were not translocated through the Sec61 translocon into the ER, we followed up on three independent approaches to generate non-imported prion proteins (Fig 1A, B, and F). First, we studied N3PrP, a full-length PrP variant equipped with an N-terminal ER signal peptide and a C-terminal GPI signal sequence that is not imported into the ER after synthesis because of a mutated ER signal peptide. The mutations prevent binding of the PrP nascent chain to the SRP (signal recognition particle) and/or gating and thereby ER import (Kim et al, 2002). Interestingly, N3PrP was not only present in the cytosol of human neuroblastoma SH-SY5Y cells but also in the nucleus (Fig 1A). Second, to study the fate of PrP under conditions of blocked ER import, we reversibly inhibited co-translational translocation of secretory proteins through the Sec61 translocon by Apratoxin S9, which is the most potent synthetic analogue of the cyclic depsipeptide natural products of the Apratoxin class (Liu et al, 2009; Chen et al, 2014; Itskanov et al, 2023). We expressed PrP lacking the C-terminal GPI-anchor signal sequence (PrPΔGPI) since the ER import of PrPΔGPI is comparable to that of GPI-anchored PrP, but PrPΔGPI is secreted (Winklhofer et al, 2003; Rambold et al, 2006; Campana et al, 2007). Thus, by using this mutant we can make sure that cytosolic localization of PrPΔGPI is a consequence of its impaired ER import and not of PrP misfolding in the ER lumen (Satpute-Krishnan et al, 2014), nor endocytosis of PrP from the plasma membrane (Sunyach et al, 2003). Fluorescence microscopy indicated the localization of PrPΔGPI within the secretory pathway of control cells, as expected for a secreted protein (Figs 1B and S1C). In cells treated with Apratoxin S9 for 4 h, PrPΔGPI was mostly detected in the nucleus (Fig 1B). To facilitate the microscopic analysis, we equipped PrPΔGPI with a C-terminal GFP and imaged mouse primary cortical neurons transiently expressing PrPΔGPI-GFP +/− Apratoxin S9 treatment to ensure that the transport of non-ER-imported PrP from the cytosol to the nucleus is not a specific feature of established cell lines. In line with our findings in established cell lines, non-ER-imported PrP was targeted to the nucleus in primary neurons (Fig 1C). Next, we generated mouse neuroblastoma N2a cell lines stably expressing PrPΔGPI-GFP. The activity of Apratoxin S9 to potently inhibit ER import was validated by immunoblotting of cell lysates. The upper band, representing N-linked glycosylated PrPΔGPI-GFP within the secretory pathway, was reduced remarkably in cells exposed to Apratoxin S9 for 4 h. In addition, the amount of PrPΔGPI-GFP in conditioned media decreased under these conditions (Fig 1D). Fluorescence microscopy analysis confirmed that similarly to PrPΔGPI, PrPΔGPI-GFP was transported from the cytosol to the nucleus after inhibiting ER import by Apratoxin S9 (Fig 1E).

Previous work revealed that ER stress results in an overall reduced rate of ER import of secretory proteins (Kang et al, 2006; Oyadomari et al, 2006; Hegde & Kang, 2008). Specifically, PrP has been shown to be co-translocationally targeted to proteasomal degradation during acute ER stress (Kang et al, 2006; Orsi et al, 2006). However, in these studies, the stabilization of PrP after proteasomal inhibition was only shown by Western blotting without addressing the cellular compartment of PrP accumulation. To assess the cellular localization of non-ER-imported PrP in a third approach, we made use of ER stress and analyzed transiently transfected cells treated for 4 h with Thapsigargin, a non-competitive inhibitor of the sarco/endoplasmic reticulum Ca²⁺ ATPase (Fig 1F). Fluorescence microscopy revealed that a fraction of PrPΔGPI-GFP was not imported into the ER and targeted to the nucleus in Thapsigargin-treated cells (Fig 1F, middle and right panels). In sum, our three-pronged approach revealed that non-ER-imported PrP, because of inefficient or impaired ER import during stress, is targeted at the nucleus.

VCP/p97 and importin-ß are required for nuclear targeting of PrP

The preceding findings raised the question of the mechanism underlying the transport of cytosolic PrP into the nucleus. We first investigated whether the nuclear import of PrP is an active process mediated by nuclear transporters. Cells transiently expressing PrPΔGPI were treated with Apratoxin S9 in combination with different importin inhibitors specific for either importin-α or importin-ß. In the presence of Importazole, an inhibitor of importin-ß, PrPΔGPI accumulated in the perinuclear region (Figs 2A and S1A). The nuclear translocation of PrPΔGPI was minimally affected by Ivermectin, which inhibits importin-α (Fig S1A). A quantitative analysis in cells stably expressing PrPΔGPI-GFP confirmed that in the presence of Importazole, the nuclear import of PrPΔGPI-GFP in Apratoxin S9-treated cells was significantly impaired (Fig 2A). Our findings are consistent with the finding that the N-terminal domain of PrP contains a cryptic NLS (Gu et al, 2003). Indeed, mutating this NLS significantly decreased nuclear localization of PrPΔNLSΔGPI in Apratoxin S9-treated cells (Fig S1B).

In the ERAD pathway, VCP/p97 is required for the retrograde translocation of misfolded secretory proteins from the ER and their targeting to cytosolic proteasomes (Ye et al, 2003; Neuber et al, 2005; Schuberth & Buchberger, 2005). To study whether VCP/p97 is involved in nuclear targeting of PrP after an aborted ER import, we treated cells transiently expressing PrPΔGPI-GFP with Apratoxin S9 and VCP/p97 inhibitors. Inhibiting VCP/p97 by three different inhibitors, DBeQ, CB-5083, or NMS-873, decreased the nuclear localization of PrP and increased the fraction of PrP in the cytosol (Fig S1C). A quantitative analysis using the stably PrPΔGPI-GFP-expressing cell line verified that nuclear targeting of non-ER-imported PrP in DBeQ-treated cells was significantly reduced (Fig 2B). To analyze the role of VCP/p97 in more detail, we immunoprecipitated VCP/p97 under non-denaturing conditions to maintain protein-protein interactions and then analyzed co-purifying proteins for the presence of PrPΔGPI by Western blotting. Since the interaction of VCP/p97 with its client proteins is only transient, we included TrapVCP in our experiments, which is a VCP/p97 mutant (E578Q) that does not release clients after binding (Hulsmann et al, 2018). Immunoblots of the input lysates confirmed that Apratoxin S9 efficiently blocked ER import. The upper band of PrPΔGPI, representing glycosylated species in the secretory pathway, was absent in lysates prepared from Apratoxin S9-treated cells (Fig 2C). An interaction of PrPΔGPI with WT VCP/p97 could not be detected by this approach, neither in control nor in Apratoxin S9-treated cells. However, PrPΔGPI co-purified with TrapVCP in Apratoxin S9-treated cells (Fig 2C). Of note, TrapVCP did not interact with PrP in control cells without Apratoxin S9, revealing that the interaction was specific for non-ER-imported PrP. Moreover, VCP/p97 interacted specifically with cytosolic PrP and not with nuclear PrP (Fig 2D). We were also able to detect a ternary complex of VCP/p97, importin-β, and PrP, suggesting a hand-over mechanism of PrP from VCP/p97 to importin-β. Mutating the cryptic NLS of PrP abolished recruitment of importin-β to the complex (Fig S1D). Taken together, nuclear targeting of cytosolic PrP after a failed ER import depends not only on importin-ß but also on VCP/p97.

The cytosolic interaction of PrP with VCP/p97 and PrP nuclear targeting are independent of ubiquitination

Ubiquitination of client proteins has been shown to mediate their interaction with VCP/p97. However, recent research has revealed that VCP is capable of interacting with substrates directly, independent of ubiquitination or adaptor proteins (van den Boom et al, 2023; Weith et al, 2018). To explore the role of ubiquitination in the interaction of non-ER-imported PrP with VCP/p97 and its nuclear targeting, cells transiently expressing PrPΔGPI-GFP were treated with Apratoxin S9 in combination with an inhibitor of the E1 ubiquitin-activating enzyme (TAK-243). Immunoblotting of whole cell lysates showed that TAK-243 efficiently inhibited ubiquitination of proteins (Fig 3A). In parallel, we immunoprecipitated TrapVCP and probed the immunopellet for PrP. Notably, non-ER-imported PrPΔGPI-GFP co-purified with TrapVCP also in TAK-243-treated cells, indicating that the PrP-VCP/p97 interaction was independent of ubiquitination (Fig 3B). Likewise, ubiquitination was not required for nuclear targeting since PrPΔGPI-GFP localized to the nucleus in cells treated with Apratoxin S9 and TAK-243 (Fig 3C). However, in the presence of TAK-243, non-ER-imported PrP was stabilized, indicating that ubiquitination was required for proteasomal degradation of PrP (Fig 3D). Importantly, we also observed nuclear accumulation of WT PrP in cells treated with Apratoxin S9 and TAK-243. This demonstrates that our findings are also relevant to untagged, GPI-anchored PrP^C (Fig S2A).

It was shown recently that the flexible N-terminal tail of VCP/p97 mediates the interaction with non-ubiquitinated substrates (van den Boom et al, 2023). Consequently, we analyzed whether the N-terminal tail of VCP/p97 is required for an interaction with non-ER-imported PrP. To this end, we immunoprecipitated a TrapVCP variant lacking the flexible N-terminal domain (Δ1-207) under non-denaturing conditions to maintain protein-protein interactions and then analyzed co-purifying proteins for the presence of PrPΔGPI-GFP by Western blotting. Similar to PrPΔGPI (Fig 2C), full-length TrapVCP interacted with PrPΔGPI-GFP in Apratoxin S9-treated cells. Remarkably, PrP did not co-precipitate with Δ1-207TrapVCP, suggesting that the interaction with non-ER-imported PrP is mediated by the flexible N-terminal tail of VCP/p97 (Fig 3E).

BAG6 and RNF126 have been described previously to interact with PrP after its aborted ER import (Hessa et al, 2011; Rodrigo-Brenni et al, 2014). To test whether these factors are implicated in nuclear targeting of non-ER-imported PrP, we down-regulated their expression by siRNAs in Apratoxin S9-treated cells transiently expressing PrPΔGPI. Silencing of neither BAG6 nor RNF126 impeded nuclear targeting of PrP (Fig S2B). As a conclusion, the interaction of VCP/p97 with PrP after its failed ER import is independent of ubiquitination but requires the flexible N-terminal tail of VCP/p97. Ubiquitination is also dispensable for nuclear targeting of non-ER-imported PrP; however, ubiquitination of PrP is required for proteasomal degradation.

VCP/p97 prevents aggregation of PrP in vitro

What is the role of VCP/p97 in nuclear targeting of PrP? We hypothesized that VCP/p97 prevents PrP aggregation in the cytosol, thereby facilitating importin-β-mediated nuclear import. To study a possible anti-aggregation activity of VCP/p97, we established an in vitro aggregation assay with purified components. This assay is based on a recombinant MBP-PrP-GFP fusion protein. The N-terminal MBP (maltose-binding protein) keeps PrP soluble during purification and can be cleaved off by TEV (tobacco etch virus) protease to initiate phase transition of PrP (Fig 3F) (Kamps et al, 2021). When MBP is cleaved off, PrP-GFP is no longer soluble in physiological buffer (10 mM Tris–pH 7.4, 150 mM NaCl) and rapidly aggregates (Fig 3F, row 1). Aggregation of PrP-GFP after TEV cleavage was prevented in the presence of recombinant VCP/p97 (Fig 3F, row 2). Upon addition of ATP to the PrP-VCP/p97 complex, PrP-GFP formed aggregates, suggesting that the anti-aggregation activity of VCP/p97 was because of a specific interaction of VCP/p97 with PrP (Fig 3F, row 2). In line with this notion, the VCP/p97 inhibitor NMS-873, which inhibited nuclear import of PrP in cells (Fig S1C), also induced PrP-GFP aggregation by releasing PrP-GFP from its binding to VCP/p97 (Fig 3F, row 3). A titration analysis with increasing VCP/p97 concentrations showed that equimolar amounts of VCP/p97 are required to prevent aggregation of PrP (Fig S3A, upper panels).

To probe for conformational differences between VCP/p97-bound PrP-GFP and aggregated PrP-GFP, we performed limited proteolysis experiments. In complex with VCP/p97, PrP-GFP was more sensitive to proteolytic digestion compared to aggregated PrP (PrP + BSA) (Fig S3B). The in vitro approaches indicate that VCP/p97 acted as a holdase to prevent a conformational transition of soluble PrP-GFP into aggregates. These results also confirmed the cell culture data that the interaction of non-ER-imported PrP with VCP/p97 is independent of ubiquitination.

PrP forms partially PK-resistant aggregates in the cytosol but not in the nucleus

During our study, we noticed that nuclear N3PrP decreased over time in contrast to cytosolic N3PrP (Fig 4A). An increase in cytosolic N3PrP was seen already at 19 h after transfection, whereas the nuclear fraction of N3PrP steadily decreased and was barely detectable at 45 h (Fig S2C, left panel). Moreover, cytosolic N3PrP formed detergent-insoluble aggregates (Fig S2C, right panel). To follow up on this observation, we analyzed the degradation of PrP in the cytosol and in the nucleus by treating cells transiently expressing N3PrP with cycloheximide (CHX) to inhibit protein translation. At the beginning of the CHX chase, N3PrP was detected both in the nucleus and in the cytosol. After 2 h of CHX treatment, the nuclear fraction of N3PrP almost disappeared, whereas N3PrP in the cytosol was still present 6 h after inhibiting protein translation (Figs 4B and S2D). The increased stability of N3PrP may indicate a misfolded conformation of N3PrP in the cytosol, which is resistant to degradation. To analyze possible differences in the biochemical and biophysical properties of PrP in the two different cellular compartments, we targeted PrP-GFP either to the nucleus or the cytosol by replacing the ER signal peptide with a NLS or nuclear export signal (NES). Fluorescence microscopy confirmed that NES-PrP-GFP and NLS-PrP-GFP are almost exclusively located in the cytosol and nucleus, respectively (Fig 4C, left panel). To probe for conformational differences, we recorded FRAP to quantify protein dynamics of NLS-PrP-GFP and NES-PrP-GFP in living cells. Fluorescence recovery of the NES-PrP-GFP assemblies in the cytosol was clearly reduced in comparison to that of the NLS-PrP-GFP assemblies, indicating differences in their material properties (Fig 4C, right panel).

As another approach to study conformational differences, we performed limited proteolysis experiments. Whole-cell lysates from NLS-PrP-GFP- or NES-PrP-GFP-expressing cells were treated with increasing concentrations of proteinase K (PK) before immunoblotting. At PK concentrations that completely digested NLS-PrP-GFP, the cytosolically localized NES-PrP-GFP resisted proteolytic degradation, suggesting that NES-PrP-GFP adopted a more aggregated conformation compared to NLS-PrP-GFP (Fig 4D). This analysis also revealed that the N-terminal domain of PrP is highly sensitive to proteolytic digestion, consistent with its intrinsically disordered structure.

RNA prevents phase transition of PrP into aggregates

The FRAP recordings revealed that nuclear PrP is in a more soluble conformation compared to cytosolic PrP. A major difference between the cytosolic and nuclear chemical milieu is the high content of negatively charged RNAs in the nucleus, which can buffer protein aggregation (Maharana et al, 2018). To test for a possible role of RNA in keeping PrP-GFP soluble, we used the in vitro aggregation assay described above (Fig 3F). Indeed, bulk RNAs purified from HeLa cells interfered with the transition of soluble PrP-GFP into aggregates (Fig 4E, row 2, Fig S3A, lower panels). The anti-aggregation activity was dependent on RNA polymers since PrP-GFP started to aggregate upon addition of RNase (Fig 4E, row 2). Moreover, in the presence of RNAs, PrP-GFP was more sensitive to proteolytic digestion, indicating that RNA acted as negatively charged polymeric cosolute to keep PrP-GFP in a soluble conformation (Fig S3B).

Cytosolic but not nuclear PrP causes a proteostasis decline

The FRAP and limited proteolysis experiments revealed that PrP adopts distinct conformations in the cytosol and the nucleus, respectively. We therefore wondered whether these conformational differences translate into specific physiological consequences. To monitor the cellular proteostasis capacity, we expressed a mutated version of the conformationally unstable firefly luciferase (Fluc) protein fused to eGFP. Proteotoxic stress leads to misfolding of Fluc-eGFP, resulting in decreased luciferase activity (Gupta et al, 2011; Park et al, 2013; Blumenstock et al, 2021). After co-expression of Fluc-eGFP with either NES- or NLS-PrP, the enzymatic activity of Fluc-eGFP was quantified by a luciferase assay. In cells expressing NLS-PrP, luciferase activity was not decreased compared to control cells. In contrast, a significant decline in luciferase activity was observed in cells expressing NES-PrP (Fig 4F, left panel). To specifically detect compartment-specific alterations in proteostasis, we targeted Fluc-eGFP either to the nucleus (NLS-Fluc-eGFP) or to the cytosol (NES-Fluc-eGFP). The co-expression of NLS-PrP did not affect the luciferase activity of neither cytosolic nor nuclear Fluc-eGFP. However, PrP accumulation in the cytosol upon NES-PrP expression significantly decreased the luciferase activity of both NLS-Fluc-eGFP and NES-Fluc-eGFP (Fig 4F, middle and right panels).

Proteotoxic stress disrupts nuclear translocation of non-ER-imported PrP and induces the formation of self-perpetuating PrP aggregates in the cytosol

As shown above, cytosolic PrP aggregates increased over time upon defective ER import, whereas the relative amount of nuclear PrP decreased (Fig 4A and B). This observation led us to hypothesize that the cytosolic PrP aggregates and proteotoxic stress induced by these aggregates disrupt further nuclear import of newly synthesized PrP. To test this possibility experimentally, we induced proteotoxic stress in cells expressing PrPΔGPI-GFP by transiently inhibiting the proteasome with Bortezomib (3 h pretreatment). Then, the reversible proteasome inhibitor was washed out, and ER import was inhibited by Apratoxin S9. After additional 4 h, the cells were fixed and analyzed by fluorescence microscopy. In cells treated with Apratoxin S9 only, the typical nuclear localization of non-ER-imported PrPΔGPI-GFP was observed. However, pretreatment with Bortezomib interfered with the transport of PrPΔGPI-GFP into the nucleus and induced the formation of cytosolic PrP aggregates (Fig 5A). To address possible long-term effects of transient proteotoxic stress, we induced the formation of cytosolic PrPΔGPI-GFP aggregates by Bortezomib and Apratoxin S9 (4 h) and analyzed the cells either directly or after 48 h recovery in fresh medium by Western blotting and fluorescence microscopy. The Western blot analysis of cells after 4 h Apratoxin S9 treatment revealed the typical loss of glycosylated PrP and the appearance of unglycosylated, non-ER-imported species (Fig S4). This was accompanied by the targeting of PrP to the nucleus (Fig 5B). In cells treated with both Apratoxin S9 and Bortezomib for 4 h, the non-ER-imported PrP was not targeted to the nucleus but remained in the cytosol (Fig 5B). The corresponding Western blot analysis confirmed that PrP was not imported into the ER under these conditions since the glycosylated PrP species disappeared (Fig S4).

After recovery for 48 h in fresh medium, nuclear PrP mostly disappeared from cells that had been pre-treated with Apratoxin S9 alone (Fig 5B). However, large perinuclear PrP aggregates were present in the cytosol of cells 48 h after exposure to Apratoxin S9 and the proteasomal inhibitor Bortezomib for 4 h (Fig 5B). These PrP aggregates were formed from non-ER-imported PrPΔGPI-GFP, permanently generated under physiological conditions because of the inefficient ER signal peptide of PrP (Kim et al, 2001; Rutkowski et al, 2001; Drisaldi et al, 2003; Heller et al, 2003; Heske et al, 2004; Rane et al, 2004; Hegde & Bernstein, 2006; Rambold et al, 2006; Miesbauer et al, 2009). In sum, these experiments indicated that short-term proteotoxic stress leads to the formation of self-perpetuating PrP aggregates in the cytosol that interfere with further nuclear targeting of non-ER-imported PrP.

DNA constructs and siRNA

Plasmid maintenance and amplification were carried out using Escherichia coli TOP10© (Thermo Fisher Scientific). All PrP constructs were generated by standard PCR cloning techniques and are based on the coding region of mouse PRNP (GenBank accession number M18070) modified to express PrP-L108M/V111M, allowing detection by the monoclonal antibody 3F4. All the mammalian expression plasmids were cloned into pcDNA3.1(+)-Neo vector (V79020; Invitrogen). PrPΔGPIGFP: aa 1–226 tagged with GFP at the C-terminus; PrPΔGPI: aa 1–226; NES-PrP-GFP: aa 1–226 tagged with GFP at the C-terminus and with a nuclear export signal at the N-terminus; NLS-PrP-GFP: aa 1–226 tagged with GFP at the C-terminus and with a NLS at the N-terminus; NES-PrP: aa 1–226 tagged with a nuclear export signal at the N-terminus; NLS-PrP: aa 1–226 tagged with a NLS at the N-terminus; N3PrP: mouse PrP with a mutated ER signal peptide (Kim et al, 2002); PrPΔNLSΔGPI: mouse PrP aa 1–226 with K23A, K24A, R25A, and K27A substitutions were cloned into pcDNA3.1(+)-Neo via HindIII and XhoI. The modified signaling sequences are as follows:

NLS: 5′ ATGCCACCAAAAAAAAAAAGAAAAGTT 3′

NES: 5′ ATGCTGGAACTTCTGGAGGATTTGACACTG 3′

N3 signal peptide: 5′ATGGCGAACCTTGGCGATGACCTGCTGGCCCTCTTTGTGACTATGTGGACTGATGTCGGCCTCTGC 3′

Generation of pMAL-MBP-TEV-PrP-eGFP-TEV-His₆ is described elsewhere (Kamps et al, 2021). pcDNA5FRT/TO-p97-EQ-mycStrep, pcDNA5FRT/TO-p97wt-mycStrep, and pGEX-6P-1-p97 were kindly provided by Hemmo Meyer (Ritz et al, 2011). Fluc-eGFP, NLS-Fluc-eGFP, and NES-Fluc-eGFP were kindly provided by F. Ulrich Hartl and Irina Dudanova (Gupta et al, 2011; Park et al, 2013; Blumenstock et al, 2021).

pcDNA3.1(+)-Neo was used to subclone VCP mutant lacking aa 1–207 (Δ1-207 TrapVCP) and with a strep tag at C-terminus via NheI and BamHI, VCP mutant was amplified from pcDNA5FRT/TO-p97-EQ-mycStrep using following primers:

Forward: 5′ ATATGCTAGCATGTGCAGGAAGCAGCTAGCTC 3′

Reverse: 5′ ATATGGATCCTTTTTCGAACTGCGGGTGG 3′

The following siRNA was used to knock down BAG6 and RNF126: (ambion Silencer Select pre-designed).

RNF126: 5′ GCAUCUUCGAUGACAGAGCUUTT 3′ (siRNA ID: s31185).

BAG6: 5′CAAGAGCAGUUUAAUAGCATT 3′ (siRNA ID: s15468).

Cell culture and transfection

Primary neuron culture

Primary neurons were prepared following the protocol from Life Technologies, with modifications. All media and solutions were sterile filtered using 0.22 μm membrane filter. Cortices were explanted from embryos 16–18 DPC (day post coitum), carefully washed thrice with ice cold Dissection Buffer (9.9 mM HEPES, 137 mM sodium chloride, 5.4 mM potassium chloride, 0.17 mM sodium phosphate dibasic anhydrous, 0.22 mM potassium phosphate monobasic anhydrous, 33.3 mM D-glucose, 43.8 mM sucrose, pH 7.4), and treated with 1 mg acetyl trypsin for 10 min at 37°C in the water bath. Cortices were then rinsed three times in dissection buffer, and 100 U DNase-I was added. Complete dissociation was performed manually using a series of increasingly smaller pipette tips, followed by straining through a 70 μm filter (EASYSTRAINER, Greiner Bio One) to remove larger tissue residues and diluted with neurobasal medium (with 10% FBS, and 0.5 mM glutamine). The number of cells was determined using a haemocytometer, and the neurons were plated on poly-L-lysine-coated plates at densities of 1.3 × 10⁴ and 1.3 × 10⁶ cells/well in 24-well plates and 60-mm dishes, respectively. The medium was changed to neurobasal medium supplemented with B-27 supplement and 0.5 mM glutamine after 50 min. The primary cortical neurons were cultured in a humidified incubator at 37°C with 5% CO₂. Five DIV (day in vitro) neurons were transfected using Lipofectamine 2000 following the manufacturer’s protocol and analyzed for immunofluorescence after 48 h. Animal protocols were performed in compliance with institutional and governmental regulations.

Immunofluorescence and confocal microscopy

Transfected cells were washed with PBS (Gibco) and then fixed using 4% (wt/vol) PFA for 10 min. Cells were permeabilized using 0.2% Triton X-100 in PBS for 10 min and washed with PBS. For blocking, cells were incubated for 1 h in 5% normal goat serum in PBS and incubated in 0.1–0.2% (vol/vol) primary antibody in the blocking solution overnight at 4°C. Cells were then washed with PBS and incubated with the corresponding 0.1% (vol/vol) secondary antibody in PBS for 1 h at 25°C. The secondary antibody was removed and washed away using PBS and 0.2% Tween 20 in PBS for 5 min. All coverslips were incubated in 0.1% (vol/vol) DAPI in Millipore, washed with Millipore water, and then mounted using Fluoromount-G (Invitrogen). For non-permeabilized cells, after fixation, coverslips were counterstained with DAPI and mounted.

Fluorescence images were acquired using the ELYRA PS.1 microscope equipped with an LSM880 (Carl Zeiss) and a 20x, 63x oil or 100x oil immersion objective. Super-resolution images were generated by structured illumination microscopy (SR-SIM) using 405, 488, and 561 nm widefield laser illumination. SIM confocal images were processed using the raw scale mode of ZEN2.3 software (Carl Zeiss). In detail, for each channel, five phase images in three different rotations were acquired. After SIM processing, the confocal section thickness (z-scaling) is either 0.126 μm (AlexaFluor555 and DAPI) or 0.110 μm (AlexaFluor488/GFP and DAPI) based on the optical properties of the Zeiss Plan-Apochromat 63x/NA1,4 Oil DIC M27 objective in combination with the 405, 488, and 561 nm diode laser lines. For the quantification of nuclear fluorescence intensities, confocal images were acquired with the imaging settings kept uniform among replicates.

Immunoblotting

Proteins were fractionated by SDS–PAGE and transferred to nitrocellulose or polyvinylidene difluoride membranes by electroblotting. The nitrocellulose membranes were blocked with 5% non-fat dry milk or 5% BSA in TBST (TBS containing 0.1% Tween 20) for 60 min at RT and subsequently incubated with the primary antibody diluted in blocking buffer for 16 h at 4°C. After extensive washing with TBST, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody for 60 min at RT. After washing with TBST, the antigen was detected with the ECL detection system (Promega) as specified by the manufacturer with Azure Sapphire Biomolecular Imager (Azure Biosystems).

Biochemical analyses

PNGase-F digestion and glycosylation assessment

In a six-well plate, HEK293T cells were seeded, grown overnight, and transfected with PrPΔGPI-GFP. After 16 h, the cells were treated with or without 100 nM Apratoxin S9 for 4 h. The cells were harvested (1,000g, 5 min), lysed in 1% Triton X-100 in PBS, vortexed, and incubated on ice for 10 min. The lysates were centrifuged at 18,000g for 10 min, and the resulting supernatants were treated with PNGase-F (New England Biolabs) for 3 h at 37°C. The samples were analyzed through sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS–PAGE) using a 10% SDS–PAGE gel and immunoblotted using an antibody, 3F4, against PrP.

Co-immunoprecipitation

In 6-cm dishes, HEK293T cells were seeded and co-transfected with PrPΔGPI and p97Wt-mycStrep or p97-EQ-mycStrep. 24 h post-transfection, cells were harvested (1,000g, 5 min), lysed with 500 μl lysis buffer (150 mM KCl, 5 mM MgCl₂, 50 mM Tris–HCl pH 7.4, 1% Triton X-100, 5% glycerol, 2 mM β-mercaptoethanol supplemented with Complete EDTA-free protease inhibitors and PhosSTOP, Roche), gently resuspended by pipetting, and incubated on ice for 20 min. The lysates were centrifuged at 18,000g for 15 min at 4°C, and the supernatant (input) was collected in separate tubes. MagStrep “type3” XT beads (IBA Lifesciences GmbH) pre-washed with lysis buffer were added to the supernatants and incubated for 2 h at 4°C with gentle rotation. The flowthrough was removed, and the beads were washed thrice with 500 μl lysis buffer. Finally, the beads were boiled with 50 μl 2x Laemmli sample buffer for 5 min to elute the immunocomplex. All samples were fractionated by SDS–PAGE gels and further analyzed through immunoblotting against VCP and PrP.

Luciferase assay

HEK293T cells were seeded and co-transfected with Fluc-GFP(Wt)/NLS-Fluc-GFP(Wt)/NES-Fluc-GFP(Wt) and NLS-PrPΔGPI/NES-PrPΔGPI/pcDNA3.1(+) empty vector control. 24 h post-transfection, cells were harvested (1,000g, 5 min), lysed in 100 μl 1X reporter lysis buffer (Promega), vortexed, and centrifuged (17,049g, 4°C, 15 min). The resulting supernatant (20 μl) was added into a 96-well plate in quadruplicate. Before loading the plate into the Cytation5 reader (BioTek), the injectors were washed following the manufacturer’s protocol and primed with the luciferase assay substrate (Promega) diluted in water (1:10). 100 μl of the substrate was injected per well, and the luminescence was measured at 557 nm. The empty vector controls were used to normalize all the samples within the same transfection group and plotted to calculate the fold change in luminescence of the Fluc reporter.

FRAP analyses

SH-SY5Y cells were seeded on a 35-mm IBIDI μ-Dish and transfected with NLS-PrP-GFP or NES-PrP-GFP. 24 h post-transfection, cells were imaged in Invitrogen Live Cell Imaging Solution. ZEN2.1 bleaching and region software module and Plan-Apochromat 100× numerical aperture 1.46-oil differential interference contrast M27 objective was used for imaging the cells. For each cell analyzed, two circular regions of interest were chosen. One region was bleached with 100% laser power and a pixel dwell time of 8.71 ms, with a scan time of 111.29 ms, and the other region was used as the reference signal. Relative fluorescence intensity (RFI) was calculated at time t using the following equation: RFI = I_BL(t)/I′_BL/I_Ref(t)/I′_Ref, where I_BL(t) and I_Ref(t) are the intensities measured at time t in the photobleached region and the reference region, respectively. I′_BL and I′_Ref are the intensities measured before photobleaching.

Proteinase K digestion

HEK293T cells were seeded and transfected with NLS-PrP-GFP or NES-PrP-GFP. 24 h post-transfection, cells were lysed in 200 μl lysis buffer (1% Triton X-100 in 50 mM HEPES pH 7.4, 150 mM NaCl) on ice for 15 min. The lysate was cleared at 17,000g for 10 min, and equal amounts of the supernatant (20 μl) were digested with the indicated concentrations of proteinase K for 15 min at 25°C. The reaction was stopped with 5 mM PMSF (phenylmethylsulfonyl fluoride), mixed with Laemmli sample buffer, and boiled for 1 min. Samples were analyzed by immunoblotting against a PrP antibody specific to C-terminus (POM-15) and GFP.

Expression and purification of recombinant proteins

Plasmid maintenance, bacterial expression, and purification of MBP-PrP-GFP were performed as described earlier (Kamps et al, 2021). For expression and purification of GST-tagged VCP, transformed E. coli BL21 (DE3) were grown to an OD 600 of 0.6–0.8 before induction with 0.5 mM isopropyl-β-d-thiogalactopyranoside and a change in temperature to 18°C overnight. Cells were harvested and resuspended in lysis buffer (50 mM HEPES pH 8.0, 150 mM KCl, 2 mM MgCl₂, 5% glycerol) with 20 ml/liter of culture volume, protease inhibitor cocktail, and 1 mg/ml of Lysozyme added to the suspension, stirred with a stir-bar gently for 30 min at 4°C, and lysed by a French press. After centrifugation for 60 min at 20,000g at 4°C, the lysate was cleared with a 0.8 µm filter. GST-tagged VCP/p97 was affinity-purified with a GSTTrap FF column (GE Healthcare) with a flow rate of 1–5 ml/min on an ÄKTA purification system. The column was washed with the lysis buffer and the bound proteins were eluted using the lysis buffer but with 20 mM glutathione. Adding glutathione changes the pH; buffer needs to be adjusted to pH7.4–8.0 before use. Eluted fractions were assessed for protein content by SDS–PAGE. Desired fractions were pooled, concentrated to <5 ml for injection from loop onto gel filtration column, and further purified using HiLoad 16/600 Superdex 200 pg at 1 ml/min flowrate with gel filtration buffer (50 mM HEPES pH 7.4, 150 mM KCl, 2 mM MgCl₂, 5% glycerol, 1 mM DTT). Eluted fractions were checked for protein content by SDS–PAGE. Desired fractions were pooled, concentrated to the desired concentration, flash frozen in liquid nitrogen, aliquoted, and stored at −80°C. The protein concentration was determined using the absorbance at 280 nm and the extinction coefficient of each protein by NanoDrop 2000 (Thermo Fisher Scientific).

Sample preparation for in vitro aggregation assay

Protein samples were thawed and centrifuged (20,000g, 10 min, 4°C) to remove aggregates. Afterwards, the buffer was exchanged to 10 mM Tris–pH 7.4 using Vivaspin 500 columns with 30,000 D cut off (Sartorius Stedim Biotech). The samples were centrifuged (12,000g, 7 min, 4°C) five times for complete buffer exchange, and then the protein concentration was determined by NanoDrop 2000. To induce aggregation of PrP-GFP, the buffer was supplemented with 150 mM NaCl and incubated with TEV protease for 1 h before microscopy.

Limited proteolysis of recombinant PrP-GFP

5 μM MBP-PrP-GFP in aggregation buffer was cleaved with TEV protease in the presence of BSA, GST-VCP, or bulk RNA prepared from HeLa cells. The samples were treated for 15 min at 25°C with increasing concentrations of proteinase K as indicated. After stopping the reactions with 5 mM PMSF (phenylmethylsulfonyl fluoride), the samples were analyzed by SDS–PAGE and Coomassie brilliant blue staining.

Quantification and statistical analysis

CellProfiler (https://cellprofiler.org) was used to calculate ratio of nuclear to cytosolic mean GFP intensity of each cell, by using DAPI as a marker for nuclear mask. The macro used for this analysis was “Human C-N translocation” written for automated image analysis which is available on the CellProfiler website. The in vitro samples were analyzed as described previously (Ahlers et al, 2021; Kamps et al, 2021). Statistical analyses for the luciferase assays and nuclear GFP intensity quantification were performed using one-way ANOVA (Kruskal-Wallis test with Dunn’s post-test at 95% confidence interval, * = P < 0.05) and t test (Mann-Whitney test, two-tailed at 95% confidence interval, * = P < 0.05), respectively, using the GraphPad PRISM 9.5 software.