Class-specific interactions between Sis1 J-domain protein and Hsp70 chaperone potentiate disaggregation of misfolded proteins

Abstract

Protein homeostasis is constantly being challenged with protein misfolding that leads to aggregation. Hsp70 is one of the versatile chaperones that interact with misfolded proteins and actively support their folding. Multifunctional Hsp70s are harnessed to specific roles by J-domain proteins (JDPs, also known as Hsp40s). Interaction with the J-domain of these cochaperones stimulates ATP hydrolysis in Hsp70, which stabilizes substrate binding. In eukaryotes, two classes of JDPs, Class A and Class B, engage Hsp70 in the reactivation of aggregated proteins. In most species, excluding metazoans, protein recovery also relies on an Hsp100 disaggregase. Although intensely studied, many mechanistic details of how the two JDP classes regulate protein disaggregation are still unknown. Here, we explore functional differences between the yeast Class A (Ydj1) and Class B (Sis1) JDPs at the individual stages of protein disaggregation. With real-time biochemical tools, we show that Ydj1 alone is superior to Sis1 in aggregate binding, yet it is Sis1 that recruits more Ssa1 molecules to the substrate. This advantage of Sis1 depends on its ability to bind to the EEVD motif of Hsp70, a quality specific to most of Class B JDPs. This second interaction also conditions the Hsp70-induced aggregate modification that boosts its subsequent dissolution by the Hsp104 disaggregase. Our results suggest that the Sis1-mediated chaperone assembly at the aggregate surface potentiates the entropic pulling, driven polypeptide disentanglement, while Ydj1 binding favors the refolding of the solubilized proteins. Such subspecialization of the JDPs across protein reactivation improves the robustness and efficiency of the disaggregation machinery.

Keywords: Hsp40; chaperones; protein aggregation.

Fig. 1.

Hsp70 system with Sis1 or Ydj1 shows differences in aggregate binding and disaggregation. (A) Refolding of aggregated luciferase by Ssa1 and Hsp104 in the presence of Sis1 (red) or Ydj1 (blue). (Inset) Magnification of the first 20 min of the reaction. Error bars show SD from three experiments. Luciferase activity was normalized to the native protein activity. Two-tailed t test: **P ≤ 0.01 and *P ≤ 0.05. (B, Upper) The scheme of the experiment. BLI sensor with immobilized luciferase aggregate was incubated with Ssa1 and Sis1 or Ydj1, with or without ATP, as indicated in the legend. Dashed lines indicate the start of the incubation with the chaperones and washing with the buffer without chaperones. Shown is a representative result for three repeats. (C) Western blot analysis of Ssa1 binding in the presence of Sis1 or Ydj1 to the sensor with luciferase aggregates. The experiment was performed as in B, except that at the end of the binding step the proteins dissociated from the sensor into the Laemmli buffer and were analyzed with Western blot. (Right) The bands intensities were quantified. Error bars show SD from three independent experiments. Two-tailed t test was performed: **P ≤ 0.01. (D) Binding of Sis1 (red) and Ydj1 (blue) without Ssa1 to luciferase aggregates immobilized on the BLI sensor, performed analogously as in B.

Fig. 2.

Sis1 and Ydj1 drive different modes of Ssa1 recruitment to aggregates. (A, Upper) The experimental scheme. BLI sensor with luciferase aggregates was initially incubated with or without Ydj1 (Left) or Sis1 (Right). After washing, the sensor was incubated with Ssa1 or JDP-Ssa1, as indicated in the legend. (B) Sequential incubation of luciferase aggregates, immobilized on the sensor with the indicated chaperones, was carried out analogously as in A. The effect of the initial incubation on the binding kinetics of Sis1-Ssa1 is indicated with an arrow. (C) Sensors with luciferase aggregates after glutaraldehyde cross-linking, and untreated were incubated with Ydj1-Ssa1 or Sis1-Ssa1. Dashed lines indicate the starting point of each step. The presented results are representative for three replicates.

Fig. 3.

Sis1 facilitates the initiation of Hsp104-dependent protein disaggregation. (A) BLI sensor covered with luciferase aggregates, and Sis1-Ssa1 (red) or Ydj1-Ssa1 (blue) was incubated with Hsp104. Washing step involved the buffer without chaperones. The presented results are representative for three replicates. (B) Western blot analysis of Hsp104 binding in the presence of Ssa1 and Sis1 or Ydj1 to the sensor covered with luciferase aggregates. The experiment was performed as in A, except instead of the dissociation step the proteins were removed from the sensor by boiling the sensor in Laemmli buffer, and subsequently, the Western blot was performed using anti-Hsp104 and anti-luciferase (control) antibodies. (Right) The bands intensities were quantified. Error bars show SD from three independent experiments. Two-tailed t test was performed: *P ≤ 0. 05. (C) Impact of initial incubation with the Hsp70 system on the efficacy of disaggregation of denatured luciferase. Luciferase activity was measured at the indicated time points. 104mut designates the Hsp104 D484K F508A variant. Luciferase activity was normalized to the native protein. Error bars show SD from three independent experiments. (Upper) Experimental schemes. Dashed lines indicate the starting points of the incubation steps.

Fig. 4.

Sis1 (CTDI)-Ssa1 (EEVD) interaction is crucial for the superior, aggregate-remodeling activity of Sis1-Ssa1. (A) Binding of chaperone variants (as color indicated in the legend) to luciferase aggregates immobilized on the BLI sensor, according to the scheme in Fig. 1B. (B) Binding of Sis1 E50A and Ssa1 ΔEEVD to luciferase aggregates cross-linked with glutaraldehyde (orange) and nonmodified (yellow), as on the scheme in Fig. 2C. The presented results are representative for three replicates. (C) Luciferase reactivation by the system upon sequential addition of different chaperone variants (as color indicated in the legend), according to the scheme in Fig. 3C. Luciferase activity was normalized to the native protein. Error bars show SD from three independent repeats. Dashed lines indicate the starting points of the subsequent steps of experiments.

Fig. 5.

Model of distinctive contribution of Class A and B JDPs to protein disaggregation. (A) While both Ydj1 and Sis1 bind and stimulate Hsp70 through the J-domain binding to Ssa1 NBD, in Sis1, the additional interaction with the EEVD motif expands the network of possible interactions. (B) Despite the initially a lower level of aggregate binding, with time, Sis1 enables more abundant loading of Ssa1 molecules on the aggregated substrate. Larger chaperone complexes likely potentiate entropic pulling (54), which remodels aggregates in a way that favors more abundant chaperone binding. Saturation of protein aggregate with Hsp70 enhances Hsp104 docking and disaggregation. Disaggregated polypeptides with misfolding proclivity are better substrates for Ydj1 than Sis1 (26). Ydj1 binding prevents their aggregation, which promotes their return on correct folding pathways.