The Co-Chaperone HspBP1 Is a Novel Component of Stress Granules that Regulates Their Formation

Abstract

The co-chaperone HspBP1 interacts with members of the hsp70 family, but also provides chaperone-independent functions. We report here novel biological properties of HspBP1 that are relevant to the formation of cytoplasmic stress granules (SGs). SG assembly is a conserved reaction to environmental or pathological insults and part of the cellular stress response. Our study reveals that HspBP1 (1) is an integral SG constituent, and (2) a regulator of SG assembly. Oxidative stress relocates HspBP1 to SGs, where it co-localizes with granule marker proteins and polyA-RNA. Mass spectrometry and co-immunoprecipitation identified novel HspBP1-binding partners that are critical for SG biology. Specifically, HspBP1 associates with the SG proteins G3BP1, HuR and TIA-1/TIAR. HspBP1 also interacts with polyA-RNA in vivo and binds directly RNA homopolymers in vitro. Multiple lines of evidence and single-granule analyses demonstrate that HspBP1 is crucial for SG biogenesis. Thus, HspBP1 knockdown interferes with stress-induced SG assembly. By contrast, HspBP1 overexpression promotes SG formation in the absence of stress. Notably, the hsp70-binding domains of HspBP1 regulate SG production in unstressed cells. Taken together, we identified novel HspBP1 activities that control SG formation. These features expand HspBP1's role in the cellular stress response and provide new mechanistic insights into SG biogenesis.

Keywords: HspBP1; chaperone; co-chaperone; proteostasis; stress granule; stress response.

Figure 1

In response to oxidative stress, HspBP1 concentrates in cytoplasmic SGs, but not in PBs. HeLa cells were incubated for 4 hours with vehicle (control) or DEM, fixed and processed for indirect immunofluorescence and in situ hybridization (Materials and Methods). Nuclei were demarcated with DAPI. (A) HspBP1 located together with the SG markers HuR, G3BP1 or TIA. With the antibody used here, TIA was detected in both SGs and processing bodies (PBs). (B) HspBP1 co-localized with polyA-RNA in SGs. HeLa cells were treated 4 hours with vehicle or DEM. Samples were hybridized with Cy3-labeled oligo-dT(50) and subsequently immunostained with antibodies against HspBP1. (C) HspBP1 did not concentrate in PBs under control or oxidative stress conditions. Co-immunostaining was performed with antibodies against HspBP1 and the PB marker Dcp1. Some SGs (>) and PBs (►) are marked. Scale bar is 20 μm.

Figure 2

HspBP1 associates with SG proteins. Protein complexes were immunopurified with antibodies against HspBP1 from control and DEM-treated cells. (A) Workflow of the experiment. (B) Western blotting. Samples were incubated with vehicle (−) or DEM (+) as specified. Crude extracts (Start) and immunoprecipitates (IP) were analyzed with antibodies against HuR, G3BP1, TIA-1/TIAR, hsp90 or Grp94 as indicated. Filters were stripped and reprobed with antibodies against HspBP1. ECL signals were quantified to determine the stress-dependent changes in the association of HspBP1 with different binding partners. The ratio of interacting protein/HspBP1 was normalized to control conditions. Means +SEM are shown for three independent experiments. The student’s t-test identified significant differences; * p < 0.05. No binding of HspBP1 to hsp90 or Grp94 was detected under these conditions, demonstrating the specificity of our assay. For each protein, all lanes are from the same blot.

Figure 3

Kinetics of HspBP1 recruitment to stress granules. HeLa cells were treated with vehicle (control) or exposed to DEM. After a 4-hour DEM treatment, cells were transferred to medium without DEM for recovery (bottom panels). Samples were fixed at the time point indicated, and the distribution of HspBP1 and HuR was examined by immunocytochemistry (Materials and Methods). HspBP1 remained associated with SGs during the recovery from stress. DAPI demarcated nuclei; scale bar is 20 μm.

Figure 4

Quantification of SG parameters in DEM-treated HeLa cells. Each dataset was normalized to the 4-hour DEM treatment. (A) HeLa cells were incubated with growth medium containing 2 mM DEM or vehicle for the times indicated. After the 4-hour treatment, cells were transferred to growth medium for recovery. The results are shown as average +SEM for 3 to 4 independent experiments. For each experiment, at least 44 cells were analyzed for every data point. G3BP1 and HuR served as SG markers. One-Way ANOVA with Bonferroni correction was performed for 2-hour DEM, 4-hour DEM and all stress recovery points. The results for 4-hour DEM were used as reference. * p < 0.05, ** p < 0.01, *** p < 0.001. (B) The SG association of HuR, G3BP1 and HspBP1 was measured for individual SGs. Different time points were assessed during stress and recovery. The pixel intensity/SG area was quantified for 3 to 4 independent experiments. Each dataset was normalized to the 4-hour DEM treatment. Except for overnight recovery (317 SGs), the total number of SGs examined for each time point was between 1145 and 2059. * p < 0.05. (C) The ratio HspBP1/G3BP1 was calculated for the pixel intensity/area. Individual datasets were normalized to the results for 4-hour DEM incubation. AU, arbitrary units.

Figure 5

(A) HspBP1 interacts with RNA in vivo and in vitro. PolyA-RNA binding was evaluated as described in detail [47]. HeLa cells were incubated with vehicle (EtOH, ethanol) or 2mM DEM for 4 hours and extracts were prepared. Starting material (ST), flow through (Ft), wash (W) and material eluted from oligo(dt)-cellulose (E) was analyzed by Western blotting. HspBP1 associated with polyA-RNA under control (EtOH) and stress (DEM) conditions. HuR provided a positive control. (B) Purified HspBP1 binds RNA homopolymers in vitro. Control and homopolymer resins were incubated with purified HspBP1. Starting material (St), unbound, heparin and 1 M NaCl wash fractions, as well as bound material were examined by Western blotting with antibodies against HspBP1. All lanes for each filter were on the same blot.

Figure 6

HspBP1 knockdown impairs SG formation. HeLa cells were incubated for 4 days with the transfection agent lipofectamine, mock-treated, transfected with a control plasmid (scrambled) or constructs targeting two different regions of the HspBP1 transcript (shRNA1 and shRNA2). (A) Crude cell extracts were probed for HspBP1; actin was used as loading control. Quantitative Western blotting revealed a significant depletion of HspBP1 for each of the sh-constructs. All lanes were on the same blot. Results are shown as average +SEM. Significant differences were identified with One-Way ANOVA combined with Bonferroni posthoc analysis; * p < 0.05. (B) HeLa cells were transfected with control DNA or individual HspBP1 knockdown plasmids and subsequently treated with DEM. SGs were identified with HuR. Scale bar is 20 μm. (C) Using HuR as a marker, the SG area/cell and number of SGs/cell were quantified for control and HspBP1 knockdown cells. Bar graphs depict averages +SEM. Significant differences were determined with One-Way ANOVA and Bonferroni correction, control cells served as the reference; scram., scrambled sequence. * p < 0.05.

Figure 7

Production of HspBP1 full length protein and deletion mutants. (A) Organization of the HspBP1 protein. HspBP1 domains and the regions interacting with hsc70 are shown for the full length protein [16]. The organization of deletion mutants is also depicted. (B) Western blot analysis of full length and mutant HspBP1. Crude extracts were prepared for transfected HeLa cells and probed with antibodies against HspBP1 or hsp72. Actin was used as loading control. Lipo, lipofectamine; ctl, control plasmid; full length HspBP1, ΔM (deletion of residues 154–195), ΔMC (deletion of residues 154–195 and 314–359) ΔC (deletion of residues 314–359). All lanes were on the same blot.

Figure 8

(A) HspBP1 overexpression induces the formation of SGs. HeLa cells were transiently transfected with plasmids encoding full length HspBP1 (FL). The SG marker HuR was used to monitor granule formation. All images were acquired with the same settings; signals for HspBP1 are therefore low in control cells. The position of several SGs is marked by arrowheads. Scale bar is 20 μm. (B) The effects of HspBP1 overexpression on SG area and number was quantified for a representative experiment. SG parameters were measured for 20 cells/condition, as described for Figure 6. Student’s t-test identified significant differences between controls and cells overexpressing HSPBP1. Comparisons were made within vehicle (EtOH) or DEM groups; *** p < 0.001. (C,D) Role of hsp70-interacting domains for SG production. The impact of HspBP1 deletion mutants on SG formation was analyzed for vehicle and DEM-treated cells as in part A and B. Statistical evaluation was performed with One-Way ANOVA plus Bonferroni correction, using control cells (ctl) as reference; * p < 0.05; *** p < 0.001. Scale bar is 20 μm.

Figure 9

Hsp/hsc70 inhibition alters the properties of DEM-SGs. (A) HeLa cells were incubated with 2 mM DEM for 4 hours. The vehicle DMSO or 10 µM apoptozole were added together with ethanol (EtOH) or DEM. Scale bar is 20 μm. (B,C) Quantification is shown as average + SEM for two independent experiments. Results were normalized to vehicle controls. At least 54 cells or 355 SGs were evaluated for each experiment and data point. Significant differences were identified with Student’s t-test; *** p < 0.001. AU, arbitrary units.

Figure 10

Simplified model depicting the role of HspBP1/hsp70 interactions for SG formation. Stress, such as the oxidant DEM, induces SG assembly. HspBP1, through its inhibition of hsp70, controls SG formation. Overexpression of HspBP1 triggers SG assembly, even in the absence of stress. HspBP1 knockdown or HspBP1 mutants with compromised hsp70 interaction impair SG assembly during oxidative stress. See text for details.