DnaJs are enriched in tau regulators

Abstract

The aberrant accumulation of tau protein is implicated as a pathogenic factor in many neurodegenerative diseases. Tau seeding may underlie its predictable spread in these diseases. Molecular chaperones can modulate tau pathology, but their effects have mainly been studied in isolation. This study employed a semi-high throughput assay to identify molecular chaperones influencing tau seeding using Tau RD P301S FRET Biosensor cells, which express a portion of tau containing the frontotemporal dementia-related P301S tau mutation fused to a FRET biosensor. Approximately fifty chaperones from five major families were screened using live cell imaging to monitor FRET-positive tau seeding. Among the tested chaperones, five exhibited significant effects on tau in the primary screen. Notably, three of these were from the DnaJ family. In subsequent studies, overexpression of DnaJA2, DnaJB1, and DnaJB6b resulted in significant reductions in tau levels. Knockdown experiments by shRNA revealed an inverse correlation between DnaJB1 and DnaJB6b with tau levels. DnaJB6b overexpression, specifically, reduced total tau levels in a cellular model with a pre-existing pool of tau, partially through enhanced proteasomal degradation. Further, DnaJB6b interacted with tau complexes. These findings highlight the potent chaperone activity within the DnaJ family, particularly DnaJB6b, towards tau.

1. Introduction

Tau pathology accumulation is the major driver for more than 20 neurodegenerative diseases, including Alzheimer’s disease (AD) [1–5]. Mutations in the MAPT gene, which encodes the tau protein, can lead to the development of frontotemporal dementia (FTD) [6, 7]. Tau pathology is the most common pathological entity observed in diseased brains, which develops in a distinct and predictable manner [8–10]. Recent research suggests this expected pattern is likely due to the propagation of tau deposits through a prion-like mechanism known as “tau seeding” [11–15]. However, the local accumulation of normal tau by pathogenic tau has been shown to limit this process [16, 17]. In mouse models of tauopathy, tau reduction prevents neuronal loss and mitigates tau accumulation and seeding [18–22]. Given the close association between tau pathology and disease progression, as well as the detectability of tau seeding prior to significant neuronal loss and notable pathology, there is a pressing need to identify regulators of tau seeding and accumulation [15, 23–25]. Such discoveries could pave the way for the development of novel therapeutic targets and strategies to effectively manage tau-related neurodegenerative diseases.

Molecular chaperones are a diverse set of proteins that play a crucial role in maintaining homeostasis and regulating protein quality control [26–28]. In the context of tau, discrete molecular chaperones are involved in the processing of normal and pathological tau in the brain [29–34]. However, in aging and neurodegenerative diseases, like AD, the balance of chaperone levels becomes dysregulated, potentially contributing to the initiation and progression of tau pathogenesis [35–37]. Despite the extensive network of chaperones and overlap in roles, there is still a high degree of functional diversity between and within many chaperone families[38–45]. Work from our group and others has shed light on the distinct interactions between tau and molecular chaperones from several major families, including heat shock protein (Hsp) 90s and Hsp90 cochaperones, Hsp70s, Hsp40s (commonly DnaJs), FK506-binding proteins (FKBPs), and small Hsps (sHsps). Among these, both Hsp90 and Hsp70, two highly abundant chaperones involved in protein triage and folding, are essential for tau processing and stability [30, 46–51]. Although the Hsp70/Hsp90 machinery often works in conjunction with cochaperones to recognize misfolded clients, such as tau, Hsp70 and Hsp90 can also act independently [52]. For example, Hsp90 can regulate tau alone or with cochaperones, such as Carboxy-terminus of Hsc70-interacting protein (CHIP) and the activator of Hsp90 ATPase homolog 1 (Aha1), which alter how Hsp90 modulates tau function and stability [40, 48, 50, 53–57]. While FKBP51 coordinates with Hsp90 to regulate tau, other FKBPs, like FKBP52 and FKBP12, regulate tau independently [33, 58, 59]. Hsp70 typically relies on DnaJ cochaperones for its function but can also affect tau turnover and phosphorylation in an isoform- and ATP-dependent manner [60–63]. The DnaJs, the largest chaperone family, can mitigate tau pathology through refolding, disaggregation, and degradation [29, 30, 32, 64–66]. DnaJC5, also known as cysteine string protein α (CSPα), can promote tau secretion [30, 67]. Lastly, the ATP-independent sHsps, including Hsp22 and Hsp27, act as “first responders” to protein aggregation and have been shown to decrease tau hyperphosphorylation, aggregation, and improve synaptic plasticity in vivo [38, 68–72].

While many chaperones have been previously studied for their effects on tau, individual members and their respective families have been predominantly examined in isolation, with limited comparative analysis across chaperone families. Moreover, the role of molecular chaperones in tau seeding is only beginning to be understood. The Tau RD P301S FRET Biosensor cell model has been used by several groups to study tau seeding in vitro [11, 14, 73, 74], which includes four studies using molecular chaperones [29, 45, 66, 75]. However, these studies are limited to exploring the effects of silenced molecular chaperones, using recombinant tau fibrils formed in the presence of molecular chaperones, or overexpression within a single chaperone family. The effects of high intracellular expression of chaperones from different families on tau seeding has not been investigated.

The present study aimed to directly compare the impact of a wide range of molecular chaperones on tau seeding using a semi-high throughput live-cell quantitative assay. Therefore, Tau RD P301S FRET Biosensor cells were used to assess the effect of chaperones from five major families on tau seeding in vitro. One family in particular was found to be enriched with members that can regulate tau seeding. Follow-up overexpression and knockdown studies were performed to confirm these hits and understand more about their effects on tau. Overall, a new semi-high throughput quantitative method to screen tau regulators was characterized, which led to the identification of novel targets that should be further examined in future work.

2. Results

2.1. Discrete molecular chaperones target tau seeding in vitro

Tau RD P301S FRET Biosensor cells have been widely used as an in vitro model of tau seeding [14, 15, 73, 76]. This HEK293T cell-based model stably expresses the tau repeat domain containing the disease associated mutation P301S, fused to either a cyan or yellow fluorescent protein (CFP/YFP). Under normal cellular conditions these tau proteins remain soluble with low background signal, however when templated by exogenous tau aggregates, the Tau RD-CFP and Tau RD-YFP proteins multimerize and create a fluorescence resonance energy transfer (FRET) that can be measured. This model has demonstrated high sensitivity to low amounts of tau aggregates, which is increased by utilizing lipofectamine to promote cellular uptake of preformed tau seeds [11, 66]. Tau RD P301S FRET Biosensor cells also have high selectivity to tau as neither tau monomers nor seeds from other pathological proteins result in significant FRET signal [11, 66, 77]. While there are caveats of this model [78], correlations have been made to in vivo studies [11]. Therefore, the Tau RD P301S FRET Biosensor model was selected to screen molecular chaperone regulators of tau seeding. Initially, a robust quantitative screening method was established in a 96-well plate format in which Tau RD P301S FRET Biosensor cells yielded a significant FRET signal when seeded with sonicated recombinant P301L tau fibrils, but not when exposed to vehicle alone (Fig. 1A–C). Imaging over time allows dynamic analysis of tau seeding in live cells. This is advantageous to the typical readout employed for this cell line, flow cytometry, which only allows for static timepoint measurements [11, 66, 79–81].

Figure 1. Schematic of semi-high throughput chaperone screen.

(A) Workflow of the semi-high throughput assay developed to identify molecular chaperones that alter tau seeding. In short, Tau RD P301S FRET Biosensor cells were transfected with chaperone plasmids, subcultured into 96-well plates with 3 wells per chaperone and seeded with sonicated recombinant P301L tau fibrils or vehicle control (100 mM sodium acetate pH 7.0) 48 hours after transfection. (B) Cells were imaged every 12 hours for 60 hours with 4 non-overlapping images per well. Threshold masking was used to measure the FRET intensity within the total cell area at each time point and then normalized to the final time point to calculate %FRET signal. (C) Representative images of FRET signal at each time point are shown from cells treated with vehicle or P301L tau seed. Scale bar = 20 μm. Image created with Biorender.com

Upon validation of the assay, Tau RD P301S FRET Biosensor cells were then transfected with molecular chaperones representing five distinct families and subcultured into 96-well plates, a strategy which aided in cell viability over the course of the assay. After 24 hours, the cells were seeded with sonicated recombinant P301L tau fibrils, and FRET signal was imaged every 12 hours for a total of 60 hours (Fig. S1). Total FRET signal within the cell for each timepoint was analyzed for each chaperone family by repeated measures ANOVA with a Greenhouse-Geisser correction and Dunnett’s post hoc test. A summary of statistical analysis can be found in Table S1. Across different publications, the percentage of FRET-positive cells can vary widely from 5–100% of the cells. This variation is driven by experimental conditions, such as the lipofectamine amount, seed concentration, type of seed, timing, cell density, and whether the readout is based on imaging methods or flow cytometry [79, 82]. In this chaperone screen the percentage of FRET-positive cells at 60 hours was approximately 18±4% across the entire screen. Similar to other screening protocols and for ease of comparison [83], the percentage of FRET-positive cells without chaperone was normalized to 100%. DnaJA2, DnaJB1, DnaJB6b, Hsp90α, and FKBP19 significantly reduced tau seeding (Fig. 2). This was evident from the notable decrease in FRET signal compared to their respective empty vector (EV) controls. Among the chaperone families tested, the DnaJs had the most significant overall impact on reducing tau seeding in vitro (Fig. 2A), while other chaperone families had only single members that reached significance, such as Hsp90α (Fig. 2B) and FKBP19 (Fig. 2C). None of the members from the Hsp70 (Fig. 2D) or sHsp (Fig. 2E) chaperone families had a significant effect on tau seeding in this assay.

Figure 2. Discrete molecular chaperones alter tau seeding in vitro.

Using the semi-high throughput Tau RD P301S FRET biosensor assay, 49 molecular chaperones from five chaperone families were screened for their effects on tau seeding. For each chaperone, FRET intensity within the total cell area at 60 hours was normalized to EV control to calculate the relative %FRET signal. Bar graphs from the 60-hour timepoint show the average of 2 independent experiments as %FRET ± S.E.M. for chaperone members of the (A) DnaJ family, (B) Hsp90 and Hsp90 cochaperone families, (C) FKBP family, (D) Hsp70 family, and (E) sHsp family compared to their EV control, respectively. Data analysis was performed by repeated measures ANOVA with a Greenhouse-Geisser correction over the course of the experiment (Fig. S1) across each family of chaperones, except for Hsp90 and Hsp90 cochaperones, which were combined. The 60-hour timepoint from Fig. S1 is displayed here for simplicity. Significance indicated as follows: *p<0.05, **p<0.01, ***p<0.001. Representative images of select chaperones found to be significant in the post hoc analysis are shown. Scale bar = 20 μm.

2.2. DnaJB1 and DnaJB6b inversely correlate with tau levels

To investigate whether changes in tau-seeded aggregation could be attributed to alterations in overall tau levels, tau overexpression in HEK293T cells was utilized to measure tau protein under non-aggregation-inducing conditions in the presence or absence of the chaperones that significantly impacted tau seeding in the FRET screen. Specifically, P301L tau was co-transfected with DnaJA2, DnaJB1, DnaJB6b, Hsp90α, and FKBP19. Following 48 hours of co-expression, cells were harvested and assessed for changes in total tau levels by western blot. Remarkably, overexpression of all tested DnaJ chaperones resulted in a robust decrease in total tau, by one-way ANOVA with Dunnett’s post hoc (Fig. 3A–B). Conversely, overexpression of Hsp90α and FKBP19 did not yield any significant changes (Fig. 3A–B). These findings suggest that DnaJ chaperones may be enriched in members that potently reduce tau protein levels, even under the condition of tau overexpression, which can contribute to the formation of soluble, pathogenic forms of tau [84].

Figure 3. Overexpression of DnaJs significantly affect intracellular tau levels.

(A) Representative western blot images of cell lysates from HEK293T cells co-transfected with P301L tau and Hsp90α, FKBP19, DnaJA2, DnaJB1, DnaJB6b, or EV control for 48 hours, prior to harvesting for western blot analysis. (B) Quantification of total tau relative to GAPDH normalized to EV control is shown from three independent experiments ± S.E.M. A one-way ANOVA with Dunnett’s post hoc test was used to identify significance. Significance indicated as follows: **p<0.01.

To further corroborate the role of DnaJA2, DnaJB1, and DnaJB6b as regulators of tau, knockdown experiments were conducted to examine whether reducing the expression of these molecular chaperones would increase tau levels. HEK293T cells were transfected with either DnaJA2, DnaJB1, DnaJB6b, or GFP-targeting (control) shRNA and subsequently transfected with P301L tau the next day. Cells were harvested 48 hours after tau transfection and assessed for changes in total tau levels by western blot. Analysis using one-way ANOVA with Dunnett’s post hoc revealed that the ablation of DnaJB1 and DnaJB6b resulted in a significant increase in total tau. In contrast, DnaJA2 knockdown did not exhibit a significant effect (Fig. 4A–B). Importantly, these observations were not influenced by the efficiency of knockdown among these chaperones (Fig. 4C–E). Collectively, these findings provide compelling evidence supporting an inverse correlation between the levels of DnaJB1 and DnaJB6b with tau levels in vitro.

Figure 4. Knockdown of DnaJB1 or DnaJB6b significantly increases total tau levels.

(A) Representative western blot of cell lysates from HEK293T cells transfected with P301L tau and shDnaJA2, shDnaJB1, shDnaJB6, or shGFP control, as indicated. (B) Quantification of total tau relative to GAPDH normalized to shGFP control is shown from two independent experiments ± S.E.M. A one-way ANOVA with Dunnett’s post hoc test was used to identify significance. Quantification of knockdown was confirmed for (C) DnaJA2, (D) DnaJB1 and (E) DnaJB6b by unpaired t-test. Significance indicated as follows: *p<0.05, **p<0.01.

2.3. DnaJB6b mitigates tau accumulation in aggressive cell models

To gain further insights into the effects of DnaJB1 and DnaJB6b on tau, a more aggressive model of tau accumulation was selected that uses a tetracycline-inducible system to express wild-type (WT) tau or tau containing the FTD-associated mutations, P301L and ΔK280, in HEK293 (iHEK) cells [34, 85, 86]. In these assays, tau expression was induced for 48 hours to create a preexisting pool of tau with a subset of aggregated tau. Then, cells were transfected with DnaJB1, DnaJB6b, or EV control. Cell lysates were collected 48 hours after transfection for western blot analysis. Notably, the overexpression of DnaJB6b, but not DnaJB1, led to a significant reduction in total tau levels in all three iHEK cell lines—iHEK P301L tau (Fig. 5A–B), iHEK WT tau (Fig. 5C–D), and iHEK ΔK280 tau (Fig. 5E–F). DnaJB6b, but not DnaJB1, also reduced the total tau recovered in the low-speed pellet fraction (Fig. S2). To start to understand how DnaJB6b lowers tau levels, the iHEK P301L tau model described above was treated with a 26S proteasome inhibitor, Peptide aldehyde (PSI), a 20S proteasome inhibitor, bortezomib (BTZ), or the lysosomal inhibitor, leupeptin (Leup). This assay revealed that proteasome inhibition counteracted the tau reductions by DnaJB6b (Fig. 6). Overall, these results suggest that DnaJB6b exerts a regulatory effect on tau accumulation across different tau species likely through proteasomal degradation.

Figure 5. DnaJB6b overexpression mitigates tau levels in iHEK P301L, WT, and ΔK280 tau cells.

Tau expression was induced by tetracycline for 48 hours in iHEK P301L, WT, and ΔK280 tau cells, followed by transfection of DnaJB1, DnaJB6b, or EV for 48 hours, prior to harvesting for western blot analysis. Representative western blot and corresponding quantification of cell lysates from iHEK (A-B) P301L, (C-D) WT and (D-E) ΔK280 tau cells expressing DnaJB1, DnaJB6b, or EV control. Quantification was performed for total tau relative to GAPDH normalized to EV control from two independent experiments ± S.E.M. One-way ANOVA with Dunnett’s post hoc test was used to identify significance. Significance indicated as follows: **p<0.01, ***p<0.001.

Figure 6. Inhibition of the proteasome counteracts the reduction of tau by DnaJB6b.

(A) Representative western blot from iHEK P301L tau cells induced by tetracycline for 48 hours, followed by transfection of DnaJB6b or EV. Twenty-four hours after transfection, cells were treated with DMSO control, PSI, Bortezomib (BTZ), or Leupeptin (Leup) for 16 hours prior to harvesting cells for western blot analysis. (B) Quantification of total tau relative to GAPDH normalized to respective EV treatment controls is shown from two independent experiments ± S.E.M. Two-way ANOVA with Šídák’s post hoc test was used to identify significance. Significance indicated as follows: *p<0.05.

2.4. DnaJB6b complexes with tau

The interaction between DnaJB6b and tau has not yet been described, so we examined whether these proteins can form a complex in a cellular environment. Specifically, iHEK P301L tau cells were treated with tetracycline for 48 hours, followed by transfection with FLAG-tagged DnaJB6b or EV. Co-immunoprecipitation for FLAG was performed 48 hours after transfection to assess the association between tau and DnaJB6b. The results revealed the formation of complexes between tau and DnaJB6b (Fig. 7), indicating either a direct protein-protein interaction or a stable multi-protein complex. This suggests DnaJB6b may play a crucial role in directly regulating tau accumulation.

Figure 7. DnaJB6b complexes with P301L tau.

Tau expression was induced by tetracycline for 48 hours in iHEK P301L tau cells, followed by transfection of FLAG-DnaJB6b, FLAG-EV, or EV for 48 hours, prior to harvesting. DnaJB6b and DnaJB6b interacting proteins were immunoprecipitated using FLAG-tag conjugated beads and analyzed by western blot as indicated. Representative blot represents results from two independent experiments.

3. Discussion

This study investigated the impact of chaperone overexpression on tau seeding using a semi-high throughput quantitative assay. Through the systematic screening of molecular chaperones across five major families, our findings revealed significant effects on tau seeding by five select chaperones, three of which belong to the DnaJs. In follow up studies, the overexpression of DnaJA2, DnaJB1, and DnaJB6b significantly reduced tau. However, only the knockdown of DnaJB1 and DnaJB6b significantly increased tau levels. This suggests an inverse correlation between the expression of DnaJB1 and DnaJB6b with tau. When tested in a more aggressive tau model, the overexpression of DnaJB6b, but not DnaJB1, led to a significant reduction in total tau. The effect of DnaJB6b on tau levels may rely on proteasomal degradation. Importantly, tau co-immunoprecipitated with DnaJB6b, which suggests DnaJB6b may affect tau through direct interactions. Overall, this work highlights DnaJB6b as a potent tau regulator.

It is important to acknowledge that the use of Tau RD P301S FRET Biosensor cells, like many models of tau pathology, has its limitations and caveats. Here, liposomes were utilized to enhance the seed uptake efficiency, but it is important to note that tau seeds can be endocytosed in the absence of liposomes by micropinocytosis in an heparan sulfate proteoglycan-dependent manner [87]. This is similar to the uptake mechanisms of other aggregating proteins like SOD1, which has been shown to enter cells through micropinocytosis or endocytosis [88, 89]. FRET signal is not a direct readout of uptake but rather a downstream effect, which has been shown to correlate with the amount of cargo or seed concentration contained within the liposomes [11]. A recent study suggests that the observed FRET signal does not directly reflect templated aggregation of tau, but rather an alternative pathology [78]. However, it is worth noting that this model is still able to capture some important elements of the pathological templating and accumulation of tau [11, 79]. Additionally, our data using this model demonstrates that effects in this FRET-based model can be recapitulated in other tau models, supporting Tau RD P301S FRET Biosensor cells as a useful tool. While used in this study, it is also important to acknowledge that overexpression can lead to atypical protein-protein interactions. However, chaperone overexpression studies were complemented by knockdown studies, which provide further evidence for a connection between DnaJB1 and DnaJB6b with tau. Lastly, a limitation of cellular models is that they do not replicate all aspects of protein aggregation observed in disease, and while there are methods for identifying disease-related pathological forms of tau, complementary in vivo studies are needed to validate in vitro observations.

Our screen identified Hsp90α, FKBP19, DnaJA2, DnaJB1, and DnaJB6b as regulators of tau seeding. While Hsp90α overexpression reduced tau seeding, no changes were measured in co-expression assays. Prior work has implicated Hsp90α as a promoter of tau fibrillization [40, 48, 90]. Taken together, this suggests the action of Hsp90α on tau may be different based on the state of tau. Further studies to fully comprehend the role of Hsp90α in tau pathology are warranted but this is challenging given the numerous Hsp90α interacting proteins, which may confound interpretation [26, 91]. Like Hsp90α, FKBP19 reduced tau seeding without effects on tau accumulation. This transmembrane chaperone has high expression in secretory tissues [92], but has yet to be investigated in the context of tau or other aggregating proteins. While evidence suggests that FKBP19 may have chaperoning activity outside of its conserved enzymatic function, more studies are required to further understand its effect on tau [93].

Identifying an enrichment of tau regulators within the DnaJ family is not completely surprising, since these data complement prior work supporting an association between DnaJs and tau [94]. One of these studies reported a twofold decrease in the amount of DnaJB1 and DnaJB6b bound to tau in brain lysate from tau transgenic PS19 mice over time, while DnaJA2 showed no changes [65]. However, DnaJA2 was shown to co-localize with pathological tau in neurons and AD brain tissue [29, 32]. Recombinantly, DnaJA2 and DnaJB1 have been shown to reduce tau aggregation [29, 32], while studies using recombinant DnaJB6b in the context of tau have not yet been reported. In a FRET Tau Biosensor cell line, similar to the one used in this study, recombinant tau fibrils pre-incubated with recombinant DnaJA2 had reduced tau seeding, whereas tau fibrils pre-incubated with recombinant DnaJB1 did not [29]. Other members of the DnaJ family have been implicated in tau regulation, including DnaJC5, which is involved in tau secretion [30], and DnaJC7, which binds multiple forms of pathological and WT tau with high affinity [65, 66]. In the same study, Tau RD P301S Biosensor cells lacking DnaJC7 or DnaJB6 exhibited a significant increase in seeding [65, 66]. However, no changes in tau seeding were found in our study with the overexpression of DnaJC5 or DnaJC7 [65, 66]. Previous research has also indicated that other DnaJs, like DnaJA2 and DnaJB1, can directly interact with tau [29, 32]. A recent review highlights differences in the J-domain of DnaJ proteins that alters the coordination with Hsp70 and may drive the varied interactions with tau between the classes of DnaJ chaperones [94]. While both DnaJA and DnaJB proteins bind to many clients, Class B DnaJs have been shown to bind tau fibrils and seeds, while Class A DnaJ proteins can bind to monomeric tau as a substrate [31, 64]. In addition, Class B DnaJs have been shown to be capable of promoting tau disaggregation as part of the Hsp70 disaggregation machinery, while Class A DnaJs do not promote disaggregation [31]. Remarkably the most significant regulator of tau accumulation in this study, DnaJB6b, has only recently been described to have a potential role in regulating tau, where knockdown of DnaJB6 increased tau seeding in a cell model similar to the one we used here [66].

DnaJB6b has been reported to counteract the aggregation of several amyloidogenic proteins, including polyQ Huntingtin, polyQ ataxin, amyloid-β, α-synuclein, and TDP-43 [95–98]. Our results show DnaJB6b forms complexes with tau, suggesting that like other DnaJ family members, DnaJB6b may also engage in direct interactions with tau. This is in line with a recent report showing DnaJB6 as a tau aggregate interactor [66] and is further complemented by prior work demonstrating DnaJB6b can directly bind other neurodegenerative disease-related proteins, including amyloid-β [97]. Moreover, recent evidence indicates an inverse correlation between DnaJB6 expression and late-onset AD [37], but further studies are needed to confirm the direct interaction between DnaJB6b and tau to understand its role in neurodegenerative disease. Our results indicate that DnaJB6b may promote tau turnover through the proteasome. These data are consistent with DnaJB6b’s effect on polyQ degradation, where turnover of pathogenic proteins depends on proteasomal degradation through Hsp70 [99]. DnaJB6b may not solely rely on the proteasome to mitigate the accumulation of client proteins. In regards to seeding, questions still remain as to how DnaJB6b can reduce tau aggregation after the addition of exogenous seeds. Does DnaJB6b protect against cytoplasmic mislocalization of endogenous tau and the early signs of apoptosis similar to SH-SY5Y cells induced with SOD-1 oligomers [89]? Does DnaJB6b use the Hsp70 chaperone machinery to disaggregate tau similar to other Class B DnaJs [31]? Does DnaJB6b somehow reduce the uptake or release of seeds? It is possible DnaJB6b is the most potent regulator of tau because it synergistically uses multiple mechanisms of action. Additional studies are warranted to identify the complementary pathways DnaJB6b uses to reduce tau.

4. Conclusion

Overall, our study demonstrates the significant impact of molecular chaperones, particularly members of the DnaJ family, on tau seeding and accumulation. DnaJA2, DnaJB1, DnaJB6b, Hsp90α, and FKBP19 were identified as regulators of tau seeding. Unlike Hsp90α and FKBP19, the DnaJ proteins were found to affect intracellular tau levels, with DnaJB6b having the most potent effects. However, further in vivo studies are necessary to confirm these findings and validate DnaJB6b as a viable target for tauopathies.

5. Materials and Method

5.1. Molecular Cloning

FLAG-Hsp90α and FLAG- Hsp90β plasmids were provided by Dr. Leonard Neckers (National Institutes of Health, Rockville, MD). BiP/Grp78 pCMV-Myc-KDEL-wt was a gift from Ron Prywes (Addgene plasmid #27164; http://n2t.net/addgene:27164; RRID:Addgene_27164). The plasmids encoded in the pcDNA4/FRT/TO/V5 backbone include HSPA14 (Addgene plasmid #19515; http://n2t.net/addgene:19515; RRID:Addgene_19515, HSPA6 (Addgene plasmid #19513; http://n2t.net/addgene:19513; RRID:Addgene_19513), HSPA1A (Addgene plasmid #19510; http://n2t.net/addgene:19510; RRID:Addgene_19510), DNAJA2 (Addgene plasmid #19519; http://n2t.net/addgene:19519; RRID:Addgene_19519), DNAJA4 (Addgene plasmid #19494; http://n2t.net/addgene:19494; RRID:Addgene_19494, DNAJB1 (Addgene plasmid #19522; http://n2t.net/addgene:19522; RRID:Addgene_19522), DNAJB2A (Addgene plasmid # 19523; http://n2t.net/addgene:19523; RRID:Addgene_19523), DNAJB4 (Addgene plasmid #19526; http://n2t.net/addgene:19526; RRID:Addgene_19526), DNAJB6b (Addgene plasmid #19528; http://n2t.net/addgene:19528; RRID:Addgene_19528), DNAJB9 (Addgene plasmid #19532; http://n2t.net/addgene:19532; RRID:Addgene_19532) and were gifts from Harm Kampinga. cDNA of DnaJA2, DnaJA4, DnaJB2a, DnaJB6b, and DnaJB9 were subcloned into pCMV6 vector with an N-terminal FLAG. The plasmids encoded in the pcDNA4/FRT/TO backbone include HspB1 (Addgene plasmid #63092; http://n2t.net/addgene:63092; RRID:Addgene_63092), HspB2 (Addgene plasmid #63093; http://n2t.net/addgene:63093; RRID:Addgene_63093), HspB3 (Addgene plasmid #63094; http://n2t.net/addgene:63094; RRID:Addgene_63094), HspB4 (Addgene plasmid #63095; http://n2t.net/addgene:63095; RRID:Addgene_63095), HspB5 (Addgene plasmid #63096; http://n2t.net/addgene:63096; RRID:Addgene_63096), HspB6 (Addgene plasmid #63097; http://n2t.net/addgene:63097; RRID:Addgene_63097), HspB7 (Addgene plasmid #63098; http://n2t.net/addgene:63098; RRID:Addgene_63098), HspB8 (Addgene plasmid #63099; http://n2t.net/addgene:63099; RRID:Addgene_63099), HspB9 (Addgene plasmid #63100; http://n2t.net/addgene:63100; RRID:Addgene_63100), HspB10 (Addgene plasmid #63101; http://n2t.net/addgene:63101; RRID:Addgene_63101) and were additional gifts from Harm Kampinga. DnaJC7 and DnaJC8 pcDNA4/FRT/TO/V5 were also purchased from Addgene. Cdc37 was PCR-amplified from a human cDNA library (Invitrogen) and cloned into the pCMV6 vector. FKBP12.6 (HsCD00022455), FKBP13 (HsCD00002722), FKBP19 (HsCD00618224), FKBP22 (HsCD00616399), FKBP23 (HsCD00640901), FKBP25 (HsCD00002303), FKBP36 (HsCD00076429), FKBP133 (HsCD00294942) were subcloned from DNASU purchased donor plasmid into the pCMV6 vector. pCMV6 (Origene, #RC223397), pCMV6 with an N-terminal FLAG tag (Origene, #PCMV6XL6) and pCDNA5 FRT/TO (Invitrogen #V6520–20) empty vectors were purchased from verified sources. Hsp70 (HSPA1B), Hsc70, DnaJA1, DnaJC5, FKBP12, FKBP52, FKBP51, FKBP38, FKBP60, Grp94, CHIP, Aha1, p23, PP5C and HOP were generated by our lab. Pet28-a(+)-TEV and pCMV6 P301L 4R0N plasmids were also generated in our lab. All plasmids were propagated in DH5α competent cells (Fisher Scientific #18265017) and verified by DNA sequencing (Eurofins Genomics). A summary of these plasmids can be found in Table S2.

5.2. Generation of recombinant P301L tau seeds

Recombinant P301L tau was produced and purified, as we have previously done in our lab [45]. In summary, a glycerol stock of 4R0N P301L tau in a pET-28a(+)-TEV vector was used to inoculate a 10 mL starter culture of LB broth with kanamycin and was incubated overnight at 37°C. Cultures were grown to an OD₆₀₀ of 0.7–0.8 then induced with a final concentration of 1 mM IPTG, grown for 3 hours, and subsequently harvested. Cells were spun down at 3,500 × g for 20 min, supernatant discarded, and each pellet resuspended in 35 mL of running buffer (20 mM Tris–HCl pH 8.0, 500 mM NaCl, 10 mM Imidazole) supplemented with 1x EDTA-free protease inhibitor cocktail III (Bimake #B14002) and 1 mM PMSF (Sigma Aldrich #P7626). Resuspended pellets were frozen and stored at −80°C.

For protein purification, each cell pellet was thawed, sonicated on ice using a Sonic Dismembrator (Fisher Scientific Model 120) for 60 pulses at 85% amplitude for 5 seconds on/10 seconds off, and spun at 50,000 × g for 30 min at 4°C to isolate supernatant. The supernatant was loaded onto a standard gravity nickel column packed with Nickel Sepharose high performance resin (Cytiva #17–5268-02), washed with 35 mL of running buffer, and eluted with 25 mL of elution buffer (20 mM Tris base pH 8.0, 500 mM NaCl, 250 mM Imidazole). TEV protease (~2–4 mg produced in-house) was added directly to the elution fraction to remove N-terminal 6x His tag, which was placed into a 3,500 MW cutoff dialysis bag for 4 hours at room temperature in a 1 L beaker of TEV cleavage buffer (50 mM Tris base pH 8.0, 250 mM NaCl). Post cleavage, the solution was then dialyzed back into running buffer overnight at 4°C and another nickel column purification was performed to remove TEV protease and uncleaved protein. Finally, using a HiLoad 16/600 Superdex 200 pg column, P301L tau was separated by size exclusion chromatography in 20 mM Tris base pH 7.6, 500 mM NaCl, 0.5 mM EDTA, and 0.5 mM DTT. Protein purity was confirmed by Coomassie stained SDS-PAGE gel to be >90% pure and protein concentration was measured by BCA (Fisher Scientific #PI23225). P301L tau was diluted to 2 mg/mL, aliquoted, and flash-frozen in liquid nitrogen for storage at −80°C.

To generate recombinant P301L tau fibrils for seeding assays, 10 μM recombinant P301L tau, previously dialyzed into 100 mM sodium acetate pH 7.0 buffer overnight, was mixed with 2.5 μM low-molecular-weight heparin (MP Biomedicals #194114) and 2 mM DTT in 100 mM sodium acetate pH 7.0 buffer supplemented with 10 μM Thioflavin T (ThT). Subsequently, the mixture was pipetted into a 96-well, black, clear-bottom nonbinding surface plate (Corning #3651) at 200 μL per well, sealed, and incubated at 37°C without agitation for 120 hours. Formation of P301L tau fibrils was confirmed by an increase in ThT fluorescence. Following ThT confirmation, aggregated P301L tau from all wells was combined into a single tube, aliquoted, flash-frozen, and stored at −80°C. Prior to use in the FRET biosensor screening assay, recombinant P301L tau fibrils were sonicated using a Sonic Dismembrator for 3 pulses at 25% amplitude for 5 seconds on/5 seconds off.

5.3. Tau FRET Biosensor Cell Screening Assay

HEK293T cells stably expressing Tau RD P301S-CFP and Tau RD P301S-YFP (ATCC #CRL-3275) were subcultured into 24-well poly-L-Lysine (Sigma-Aldrich #P4707) coated plates (0.5 × 10⁶ cells per well) and co-transfected using 2.5 μL Lipofectamine 2000 (Invitrogen #11668019) per well with mKate (Evrogen #pmKate2-C) and chaperone plasmids at a 1:4 ratio of mKate to chaperone (0.2 μg mKate:0.8 μg chaperone). Twenty-four hours after transfection, cells were subcultured into 96-well, black, clear-bottom plates (Corning #3601) at a density of 0.4 × 10⁵ cells per well and allowed to recover for 24 hours, reaching about 50–60% confluency, before seeding with a final concentration of 300 nM sonicated recombinant P301L tau fibrils in 100 mM sodium acetate pH 7.0 per well using 0.25 μL Lipofectamine 2000 per well. Cells were then housed in a BioSpa8 and imaged on a Cytation 3 using brightfield, CFP-YFP FRET (excitation 445 nm/emission 542 nm—BioTek #1225110), and Texas Red (excitation 486 nm/ emission 647 nm—BioTek #1225102) LED cubes using a 10x objective (Plan Fluorite WD 2.7 NA 0.6) every 12 hours for 60 hours. Each condition was measured in triplicate, and 4 images per well were used for analysis with Gen5 Image Prime Software plus Spot Counting. Relative FRET positive intensity was calculated as the FRET positive aggregate area within total cell area (using the brightfield image) and then normalized to the EV control. mKate (Texas red) was used to confirm successful transfection.

5.4. Follow-up cell-based assays

For co-overexpression studies, HEK293T cells (ATCC #CRL-3216) were subcultured into 12-well plates (Thermo Scientific #130185) coated with poly-L-Lysine and co-transfected using 2.5 μL Lipofectamine 2000 with P301L tau and chaperone plasmids at a ratio of 1:10, tau:chaperone (0.1 μg tau: 1 μg chaperone for 12-well plates). Forty-eight hours after transfection, cells were washed with PBS and harvested for western blots.

For shRNA studies, HEK293T cells were subcultured into 12-well plates coated with poly-L-Lysine and co-transfected using with chaperone shRNA (1 μg) using 2.5 μL Lipofectamine 2000, and subsequently transfected with P301L tau (0.1 μg) twenty-four hours later. Forty-eight hours after tau transfection, cells were washed with PBS and harvested for western blots.

For the iHEK cell lines (Invitrogen #R71007), iHEK P301L, WT, or ΔK280 tau cells were subcultured into 12-well plates coated with poly-L-Lysine with complete media containing 1 μg/mL tetracycline (Sigma-Aldrich #T7660–5G) to induce tau expression. After 48 hours, media was changed to complete media without tetracycline and transfected with DnaJ chaperone and control plasmids (1 μg) using 2.5 μL Lipofectamine 2000. Forty-eight hours after transfection, cells were washed with PBS and harvested for western blots.

For proteasome and lysosome inhibition assays, iHEK P301L tau cells were subcultured into 12-well plates and treated with tetracycline, as described above. After 48 hours, the cells were transfected with DnaJ chaperone and control plasmids (1 μg). Media was replaced 24 hours after transfection with complete media containing Bortezomib (50 nM) (Selleck Chemicals #S10103), Peptide aldehyde (PSI) (10 μM) (APExBio #A1900), Leupeptin (200 nM) (Millipore #E18), or vehicle control (1% DMSO). Cells were harvested 16 hours after treatment for western blot analysis.

For co-immunoprecipitation studies, iHEK P301L tau cells were subcultured into 100 mm dishes (Sarstedt #83–3902) coated with poly-L-Lysine, with complete media containing 1 μg/mL tetracycline to induce tau expression. After 48 hours, media was changed to complete media without tetracycline and transfected with either N-terminal FLAG-tagged DnaJB6b (FLAG-DnaJB6b), N-terminal FLAG-tagged EV (FLAG-EV), or EV plasmids (12 μg) using 12.5 μL Lipofectamine 2000. Forty-eight hours following transfection, cells were harvested in IP Lysis Buffer (Thermo Scientific Pierce #PI87787) and co-immunoprecipitated with Anti-DYDDDDK Magnetic Agarose beads (Thermo Scientific Pierce #PIA36797), according to the manufacturer’s protocol, and processed by western blot.

5.5. Western blotting and antibodies

All samples, diluted into a final concentration of 1× sodium dodecyl sulfate (SDS) sample buffer (Bio-Rad #1610737) and heated at 100°C for 10 min, were loaded into Any kD SDS-PAGE gels (Bio-Rad #4569034) and transferred to PVDF membranes (MilliporeSigma #IPVH00010) followed by blocking for 1 hour in 7% milk. Rabbit anti-tau polyclonal antibody (Dako/Agilent #A0024) at 1:10,000 was used to detect total tau protein and mouse anti-GAPDH monoclonal antibody (Proteintech #1E6D9) was used at 1:1,000 to detect loading control. Mouse anti-FLAG M2 (Sigma-Aldrich #F1804) and rabbit anti-FKBP11/FKBP19 (Sigma-Aldrich #HPA041709) were used to detect overexpressed chaperones in the initial co-expression assays. Rabbit anti-DnaJA2 antibody (Thermo Scientific # PA5–63874), rabbit anti-DnaJB1 antibody (Cell Signaling #4868), and mouse anti-DnaJB6 antibody (Proteintech #66587–1) were used at 1:1,000 to detect respective DnaJs. HRP-conjugated secondary antibodies at 1:1,000 for mouse (Southern Biotech #1030–05) and rabbit (Southern Biotech #4010–05) were used for ECL visualization by the FluorChem System E (ProteinSimple # 92–14860-00).

5.6. Statistical analysis

Data distributions were first tested using the Kolmogorov–Smirnov test for normality and Bartlett’s test for homogeneity of variance, and all groups were found to have no violations. Time-course imaging data were analyzed by repeated measures ANOVA with the Greenhouse-Geisser correction and Dunnett’s post hoc tests using SPSS Statistics 28 (IBM Corp, Version 28.0. Armonk, NY) to compare the effect of chaperone overexpression within chaperone families on tau seeding, as described in the figure legends and in Table S1. In groups containing more than one control vector, which include DnaJ, Hsp90, and Hsp70 families, a repeated measures ANOVA determined the control vector data did not statistically differ across all time points, so the control vector data was grouped together to allow whole family comparisons. Statistical analysis for cell-based assays analyzed by western blot were done through one-way ANOVA with respective post hoc tests, as described in figure legends and in Table S1, using GraphPad 9 (Prism. La Jolla, CA). All graphs were generated using GraphPad 9. Significant differences of the mean values were determined at P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001).

Highlights.

• Molecular chaperones across five distinct families were assessed to determine their effect on tau seeding.

• Hsp90α, FKBP19, DnaJA2, DnaJB1, and DnaJB6b significantly reduced tau seeding.

• DnaJA2, DnaJB1, and DnaJB6b reduce intracellular tau levels.

• DnaJB1 and DnaJB6b knockdown increases tau levels.

• DnaJB6b forms complexes with tau.