Article Figures & Data
Figures
- Figure 1. Aggregation state of NM-HA in mouse neuroblastoma cell populations.
(A) Immunofluorescence staining of mouse N2a neuroblastoma cells stably expressing soluble yeast Sup35 prion domain NM tagged with the HA antibody epitope (NM-HAsol) and N2a subclones 1c, 2e, and 3b persistently producing NM-HA aggregates (NM-HAagg). NM was detected using mAb anti-HA (green), and nuclei were stained with Hoechst (blue). Scale bar: 5 μm. Note that individual cell clones differ in their respective NM-HA expression levels (Krammer et al, 2009). Cell clones had been isolated from a N2a NM-HA bulk population exposed to recombinant NM amyloid fibrils. (B) The presence of SDS-resistant NM-HA in N2a subclones was determined by a filter trap assay. NM was detected using mAb anti-HA. (C) SDD–AGE analysis of lysates from N2a cells expressing NM-HAsol and NM-HAagg. NM-HA was detected using mAb anti-HA. (D) Schematic illustration of NM-HA. A region highly abundant in Q/N at the amino-terminus is shown in blue (residues 1–38). Note that also other regions in N are enriched in Q/N. The octapepetide repeat region is depicted in green, and the carboxyterminal region of N is marked in black. Chymotrypsin preferentially cleaves at the carboxyl side of amide bonds of tyrosine, tryptophan, and phenylalanine (symbolized by arrow heads). Note that the M domain lacks these residues. Three tyrosine residues are present in the HA epitope. Putative fragments that correspond to the length of peptides identified by Western blot are indicated. (E, F) NM-HA aggregates derived from different N2a cell clones exhibit comparable chymotrypsin patterns. Lysates of N2a NM-HAsol and N2a NM-HAagg clones 1c, 2e, and 3b were subjected to increasing amounts of chymotrypsin or left untreated. Proteins were analyzed by SDS–PAGE and Western blot. NM was detected by 4A5 anti-M domain antibody raised against an epitope spanning amino acid residues 229–247 (Krammer et al, 2008b) (E) or mAb anti-HA (F) (both antibody-binding sites indicated in (D)).
- Figure 2. Interactomes of soluble and aggregated NM-HA are enriched for intrinsically disordered and nucleic acid–binding proteins.
(A) Analyzed cell lines. Biological triplicates of N2a subclones 1c, 2e, and 3b, N2a NM-HAsol and wild-type N2a cells were subjected to mass spectrometry analysis after SDS–PAGE and in-gel trypsin digestion of immunoprecipitated NM-HA using anti-HA antibodies. (B) Number of proteins found as putative interactors of soluble (NM-HAsol) and aggregated NM-HA (NM-HAagg) from cell clones 1c, 2e, and 3b combined. (C) Putative interactors of NM-HAsol and the NM-HAagg derived subclones 1c, 2e, and 3b individually. (D) Interactomes of soluble and aggregated NM-HA are enriched for intrinsically disordered proteins. Box-plot shows intrinsic protein disorder of interactors calculated using DisProt (Dunker et al, 2002). Intrinsic disorder of interactors was compared with that of a random subset of the mouse proteome. P-values are 1.7 × 10−8 and 3.48 × 10−9, respectively. Statistics were performed using the Kolmogorov–Smirnov test. Shown below is the corresponding area under the ROC curve (AUROC). (E) Interactomes of soluble and aggregated NM-HA are enriched for nucleic binding proteins. Box-plot displays the nucleic acid binding ability of interactors calculated using the scale of nonclassical RBD (Castello et al, 2012) compared with that of a random subset of the mouse proteome. P-values are 3.30 × 10−13 and 3.46 × 10−11, respectively (Kolmogorov–Smirnov test). Shown below is the AUROC. (F) Intrinsic disorder and increased nucleic binding ability are characteristics of SG proteins. Box-plot shows intrinsic protein disorder calculated (upper panel) using DisProt (Dunker et al, 2002) and nucleic acid binding ability (lower panel) calculated using the Castello scale (Castello et al, 2012) for SGs components compared with a random subset of the human proteome. (G) Interactors of NM-HAagg have a higher propensity to assemble into cytoplasmic granules compared with a random subset of the mouse proteome (Bolognesi et al, 2016). Interactors NM-HAsol versus NM-HAagg: P = 0.2973; NM-HAsol versus random proteome: P = 0.2101; NM-HAagg versus random proteome: P = 2.348 × 10−7. Statistics were performed using the Wilcoxon test.
- Figure 3. Gene Ontology enrichment analysis of NM-HA interactomes.
For all interactomes, the top 20 Gene Ontology biological process annotations were determined and combined in one list. Depicted are the percentages of genes that fall into this category compared with the respective interactome. Benjamini P-value is shown.
- Figure 4. Comparison of SGs and NM-HA interactomes.
(A) Number of SG-associated proteins identified in the NM-HA interactomes. Interactomes were compared with a list of SG interactors curated from literature (Table S1) (Jain et al, 2016; Markmiller et al, 2018). Shown is the percentage of NM-HA interacting proteins that have been reported to be SG components. (B) Number of SG components interacting with both NM-HAsol and NM-HAagg. (C) Number of SG components associated with NM-HAagg in individual subclones 1c, 2e, and 3b (Table S1). (D) Enrichment of proteins with PrlDs in NM-HAsol, NM-HAagg, and SG interactomes (Table S1) compared with the human proteome (in fold change).
- Figure S1. The N2a NM-HAagg bulk population propagates SDS-resistant NM-HA prions.
(A) The presence of SDS-resistant NM-HA polymers in the N2a NM-HAagg bulk population was determined by a filter trap assay (top panel). SDS–PAGE and Western blot analysis were performed as loading control (bottom panel). NM-HA was detected using mAb anti-HA, and actin was detected using mAb anti-actin. (B) SDD–AGE analysis of cell lysates from wild-type N2a cells, N2a cells expressing soluble NM-HA, and the NM-HAagg bulk population. NM-HA was detected using mAb anti-HA. (C) Quantitative analysis of cells bearing aggregates in the N2a NM-HAagg bulk population. Transduced cells were plated in 96-well plates, and six wells were analyzed by automated confocal microscopy. Shown is the number of cells with aggregates within the cell population expressing the transgene. Bars represent mean values ± SD.
- Figure 5. Shared components of NM-HA prions and SGs.
(A) Immunofluorescence staining of N2a NM-HAsol and N2a NM-HAagg bulk cells. NM-HA was detected using mAb anti-HA (red) and SG marker TIA-1 was detected using pAb anti-TIA-1 (green). Nuclei were stained with Hoechst (blue). Scale bar: 5 μm. (B) IP of G3BP and TIAR from lysates of wild-type N2a, N2a NM-HAsol, and N2a NM-HAagg bulk cells using mAb anti-G3BP or pAb anti-TIAR, followed by SDS–PAGE and Western blot. Total cell lysate (extract) was loaded as control. G3BP and TIAR were detected using mAb anti-G3BP and pAb anti-TIAR. NM-HA was detected using mAb anti-HA. (C) N2a NM-HAsol and N2a NM-HAagg cells were incubated with 1 μM SYTO RNASelect for 30 min and subsequently fixed with methanol, followed by immunofluorescence staining. NM was detected using mAb anti-HA (red), and RNA was visualized with SYTO RNASelect (green). Nuclei were stained with Hoechst (blue). Scale bar: 5 μm. (D) N2a NM-HAsol cells were treated with 0.5 mM sodium arsenite for 1 h to induce SGs or cells were left untreated. Immunofluorescence staining was performed using mAb anti-HA (red) and pAb anti-TIAR (green). Nuclei were stained with Hoechst (blue). Scale bar: 5 μm. (E) Protein synthesis was analyzed in N2a NM-HAsol and N2a NM-HAagg cells using the SUnSET method (Schmidt et al, 2009). Cells were incubated with puromycin for 30 min. Incorporated puromycin was detected using mAb anti-puromycin (12D10). (F) Quantitative analysis of puromycin incorporation. Bars represent mean values ± SD. Statistical analysis was performed using t test (n = 3). Significant changes are indicated by asterisks (***P ≤ 0.001). (G) Viability of N2a NM-HAsol cells and N2a NM-HAagg cells was determined by XTT tetrazolium salt assay. Bars represent mean values ± SD. Statistical analysis was performed using t test (n = 3). Changes are not significant (ns).
Source data are available for this figure.
- Figure S2. TDP-43 and FUS colocalize with NM-HA aggregates.
(A) IP of TIA-1 from lysates of wild-type N2a, N2a NM-HAsol, and N2a NM-HAagg cells using mAb anti-TIA-1, followed by SDS–PAGE and Western blot. Total cell lysate (extract) was loaded as control. TIA-1 was detected using pAb anti-TIA-1. NM-HA was detected using mAb anti-HA. (B) Immunofluorescence staining of N2a NM-HAsol and N2a NM-HAagg cells. NM was detected using mAb anti-HA (red), and endogenous FUS was detected using pAb anti-FUS (green). Nuclei were stained with Hoechst (blue). Scale bar: 5 μm. (C) Wild-type N2a, N2a NM-HAsol, and N2a NM-HAagg cells were transfected with a plasmid coding for FLAG-FUS. After 48 h, FUS was immunoprecipitated by using mAb anti-FLAG, followed by SDS–PAGE and Western blot. Total cell lysate (extract) was loaded as control. FUS was detected using mAb anti-FLAG, and NM-HA was detected using mAb anti-HA. (D) N2a NM-HAsol and N2a NM-HAagg cells were transfected with a plasmid coding for TDP-43-YFP and subjected to immunofluorescence staining after 48 h. NM was detected using mAb anti-HA (red), and TDP-43 was detected using pAb anti-TDP-43 (green). Nuclei were stained with Hoechst (blue). Scale bar: 5 μm. (E) Wild-type N2a, N2a NM-HAsol, and N2a NM-HAagg cells were transfected with a plasmid coding for TDP-43-YFP. After 48 h, TDP-43 was immunoprecipitated by using mAb anti-TDP-43, followed by SDS–PAGE and Western blot. Total cell lysate (extract) was loaded as control. TDP-43 was detected using mAb anti-TDP-43, and NM was detected using mAb anti-HA.
Source data are available for this figure.
- Figure S3. Interaction of NM-GFPagg with G3BP and p62.
(A) Bulk population of N2a cells stably expressing NM-GFP (N2a NM-GFPsol) and N2a clone stably propagating NM-GFP aggregates (N2a NM-GFPagg) (Hofmann et al, 2013). N2a NM-GFPagg cells were generated by exposing N2a NM-GFPsol cells to recombinant NM fibrils and subsequent limiting dilution cloning. Scale bar: 30 μm. (B) Validation of interaction of aggregated NM-GFP with G3BP and p62. NM-GFP in lysates of N2a NM-GFPsol or N2a NM-GFPagg cells was immunoprecipitated using GFP-trap magnetic beads. Co-immunprecipitated G3BP or p62 was detected by Western blot. Note that GFP was slightly degraded. Additional lanes were excised for presentation purposes (dashed line). (C) Control IP using unspecific IgG or anti-G3BP or anti-p62 IgGs and cell lysates of N2a NM-GFPagg. Additional lanes were excised for presentation purposes (dashed lines). (D) Pull-down of Flag-FUS transiently expressed in N2a NM-GFPsol and N2a NM-GFPagg cells. GFP-trap magnetic agarose beads were used for IP. (E) Colocalization of p62 with NM-GFP in N2a NM-GFPagg cells. N2a NM-GFPsol and N2a NM-GFPagg were stained with Hoechst and anti-p62. Scale bar: 5 μm.
- Figure S4. Association of NM-HA and NM-GFP with SGs is transient.
(A) N2a NM-HAsol cells were either left untreated or exposed to 0.5 mM arsenite for 30 min. Medium was replaced, and cells were incubated for up to 2 h posttreatment (p.t.). Fixed cells were stained using anti-HA and anti-TIAR antibodies. Scale bar: 10 μm. (B) N2a NM-GFPsol cells exposed to 0.5 mM arsenite for 30 min. Upon replacement of medium, cells were allowed to recover and fixed up to 7 h p.t. The panels on the left show representative individual cells. The panel on the right shows an overview of treated cells (scale bars: 10 and 20 μm, respectively). Arrows indicate SGs. (C) Percentage of cells of (B) with TIAR-positive SGs (only red channel). Treatment was performed in triplicates. For each time point, three images per independent experiment were analyzed. At least 400 cells per time point for each experiment were analyzed. (D) Percentage of NM-GFP expressing cells with SGs and colocalization of NM-GFP with SGs. Images of (B) were quantified. At least 300 cells per time point for each experiment were analyzed.
- Figure 6. Interactome validation of proteins involved in degradation pathways.
(A) IP of endogenous p62 from cell lysates of wild-type N2a, N2a NM-HAsol, and N2a NM-HAagg cells using mAb anti-p62, followed by SDS–PAGE and Western blot. Total cell lysate (extract) was loaded as control. p62 was detected using mAb anti-p62, and NM-HA was detected using mAb anti-HA. Pull-downs using unspecific IgG served as controls. (B) Bulk N2a NM-HAagg cells were either not transfected (for p62 detection) or transfected with constructs encoding for VCP-EGFP, Keap1-GFP, or Ubiquilin-2-FLAG and subjected to immunofluorescence staining 48 h posttransfection. NM-HA was stained using mAb anti-HA (red), p62 was detected using mAb anti-p62 (green), and FLAG was stained using mAb anti-FLAG (green). GFP is shown in green. Nuclei were stained with Hoechst (blue). Scale bar: 5 μm.
- Figure S5. N2a NM-HAsol cells were left untreated (top row, for p62 staining) or transfected with constructs coding for VCP-GFP, Keap1-GFP, and Ubiquilin-2-FLAG and subjected to immunofluorescence staining 48 h posttransfection.
NM-HA was stained using mAb anti-HA (red), p62 was detected using mAb anti-p62 (green), and FLAG was stained using mAb anti-FLAG (green). GFP is shown in green. Nuclei were stained with Hoechst (blue). Scale bar: 5 μm.
- Figure 7. Fibril-induced NM prions do not evolve from SGs.
N2a NM-HAsol cells were incubated with 5 μM recombinant NM fibrils (monomer concentration) for the indicated time points and subsequently analyzed for NM-HA aggregate induction and SGs. As a positive control, cells were incubated with 0.5 mM sodium arsenite for 1 h without addition of fibrils. Immunofluorescence staining was performed using mAb anti-HA (green) and pAb anti-TIAR (red). Nuclei were stained with Hoechst (blue). Scale bar: 10 μm.
Supplementary Materials
