Pathogenic polyglutamine tracts are potent inducers of spontaneous Sup35 and Rnq1 amyloidogenesis

Abstract

The glutamine/asparagine (Q/N)-rich yeast prion protein Sup35 has a low intrinsic propensity to spontaneously self-assemble into ordered, beta-sheet-rich amyloid fibrils. In yeast cells, de novo formation of Sup35 aggregates is greatly facilitated by high protein concentrations and the presence of preformed Q/N-rich protein aggregates that template Sup35 polymerization. Here, we have investigated whether aggregation-promoting polyglutamine (polyQ) tracts can stimulate the de novo formation of ordered Sup35 protein aggregates in the absence of Q/N-rich yeast prions. Fusion proteins with polyQ tracts of different lengths were produced and their ability to spontaneously self-assemble into amlyloid structures was analyzed using in vitro and in vivo model systems. We found that Sup35 fusions with pathogenic (>or=54 glutamines), as opposed to non-pathogenic (19 glutamines) polyQ tracts efficiently form seeding-competent protein aggregates. Strikingly, polyQ-mediated de novo assembly of Sup35 protein aggregates in yeast cells was independent of pre-existing Q/N-rich protein aggregates. This indicates that increasing the content of aggregation-promoting sequences enhances the tendency of Sup35 to spontaneously self-assemble into insoluble protein aggregates. A similar result was obtained when pathogenic polyQ tracts were linked to the yeast prion protein Rnq1, demonstrating that polyQ sequences are generic inducers of amyloidogenesis. In conclusion, long polyQ sequences are powerful molecular tools that allow the efficient production of seeding-competent amyloid structures.

Figure 1. PrD-polyQ fusion proteins form seeding-competent amyloid structures in vitro.

(A) Schematic representation of MBP fusion proteins with polyQ tracts of different lengths. (B) Time-resolved analysis of polyQ-mediated aggregation of PrD fusions by FRA. SDS-resistant protein aggregates retained on filter membranes were detected using an anti-Sup35 antibody. (C) Electron micrographs of polyQ-containing PrD protein aggregates. MBP fusions were digested with factor Xa without agitation at 37°C for 24h, stained with 1% uranyl acetate and visualised using a Philips CM100 transmission electron microscope. (D) Schematic description of an in vitro seeding assay. The formation of seeding-competent polyQ-containing PrD protein aggregates after cleavage of MBP fusion proteins with factor Xa and the subsequent conversion of HisSup35 protein from the soluble to the amyloidogenic state is shown. SDS-stable protein aggregates were monitored by FRA using specific antibodies. (E) Analysis of HisSup35 aggregate formation by FRA. MBP-PrD-polyQ fusion proteins together with MBP-HisSup35 or MBP-HisCT were incubated with factor Xa at 37°C for 24 and 48 h. Aggregates retained on filter membranes were detected using anti-His and anti-Sup35 antibodies.

Figure 2. PrD and Sup35 fusion proteins with pathogenic polyQ tracts form amyloid structures in vivo.

(A) Schematic representation of PrD- and Sup35-polyQ fusion proteins expressed in GT17 yeast cells. (B) Analysis of SDS-stable PrD and Sup35 protein aggregates with polyQ sequences by FRA. Protein aggregates retained on filter membranes were detected using anti-Sup35 and anti-polyQ antisera. (C) Immunofluorescence micrographs of GT17 yeast cells expressing PrD- and Sup35-polyQ fusion proteins. Cells were fixed, permeabilised, and probed with an anti-polyQ antibody and a Cy3-conjugated anti-rabbit IgG. Nuclei were counterstained with Hoechst 33258. Scale bar, 1µm. (D) Electron micrographs of GT17 cells expressing a Sup35Q54 fusion protein. Cells were immunogold labeled using an anti-polyQ antibody. A cytoplasmic inclusion body with aggregated Sup35Q54 protein was detected by immunogold staining. Rows of gold particles indicate the formation of Sup35Q54 fibrils.

Figure 3. Formation of PrD- and Sup35-polyQ amyloids stimulates endogenous Sup35 aggregation in yeast.

(A) Expression of polyQ-containing PrD and Sup35 fusion proteins in [psi⁻][pin⁻] GT17 yeast cells induces the spontaneous appearance of ADE+ colonies on SD-Ade plates. (B) Analysis of cell lysates of GT17 cells co-expressing the proteins PrD/Sup35, PrD/Q52 or PrD/Q90 by FRA using anti-Sup35 and anti-polyQ antisera. (C) Growth assays of [psi⁻][pin⁻] GT17 cells co-expressing the proteins PrD/Sup35, PrD/Q52 or PrD/Q90 on SD-Ade plates. (D) Analysis of endogenous Sup35 and Rnq1 protein aggregation in GT17 cells overexpressing Sup35Q54 by centrifugation assays. Proteins were detected by Western blotting using anti-Sup35 and anti-Rnq1 antibodies. P, pellet fraction; S, supernatant fraction.

Figure 4. Loss of Hsp104 activity does not prevent the formation of Sup35-polyQ aggregates in yeast.

(A) Analysis of Sup35Q54 and Sup35Q92 aggregation in OT78 cells lacking a functional Hsp104 protein (ΔHsp104) by FRA. Protein aggregates retained on filter membranes were detected using an anti-Sup35 antibody. Cell extracts prepared from GT17 cells containing a functional Hsp104 protein (control) were also examined. (B) Growth assays of GT17 and OT78 cells expressing the proteins Sup35Q54 or Sup35Q92 on SD-Ade selective plates. (C) Analysis of Sup35Q54 and endogenous Sup35 aggregation in GT17 and OT78 cells by centrifugation assay. Proteins were detected by Western blotting using an anti-Sup35 antibody. P, pellet fraction; S, supernatant fraction.

Figure 5. Sup35-polyQ aggregates are not required for propagation of endogenous Sup35 aggregates.

(A) Analysis of protein extracts prepared from 5-FOA treated yeast cells by FRA. ADE+ GT17 cells expressing the proteins Sup35Q54 or Sup35Q92 were plated onto 5-FOA medium in order to eliminate URA3 expression plasmids. SDS-stable aggregates retained on filter membranes were detected using anti-Sup35 and anti-polyQ antibodies. (B) Growth assays of 5-FOA treated GT17 cells on SD-Ade plates. 5-FOA treated yeast cells showed growth on SD-Ade plates, indicating that they contain insoluble, endogenous Sup35 aggregates. (C) Schematic representation of growth assays with 5-FOA treated yeast cells on different selective plates. Yeast cells were grown for many generations on non-selective media in order to investigate the propagation of insoluble Sup35 aggregates. (D) Analysis of Sup35 aggregation in 5-FOA treated yeast cells lacking Sup35-polyQ fusion proteins (ΔSup35Q54 and ΔSup35Q92) by centrifugation assays. P, pellet fraction; S, supernatant fraction. (E) GdnHCl treatment of 5-FOA treated yeast cells (P1) lacking the proteins Sup35Q54 and Sup35Q92. After GdnHCl treatment P1 cells were unable to grow on SD-Ade plates, indicating that they do not contain insoluble Sup35 aggregates.

Figure 6. Aggregates of Rnq1-polyQ fusion proteins stimulate the polymerization of endogenous Sup35 protein.

(A) Cartoon illustrating that formation of seeding-competent Rnq1-polyQ aggregates, which promote polymerization of endogenous Rnq1 and Sup35. Spontaneous assembly of Sup35 protein aggregates is monitored by growth assays on SD-Ade plates. (B) Schematic representation of Rnq1-polyQ fusion proteins. (C) Analysis of cell extracts prepared from yeast strains expressing polyQ-containing Rnq1 proteins by FRA. (D) Growth assays on SD-Ade selective plates. GT17 [pYex2T-Sup35] yeast cells overexpressing the fusion proteins Rnq1Q54 or Rnq1Q91 form ADE+ colonies on SD-Ade plates.

Figure 7. A model for polyQ- and template-mediated Sup35 aggregation in yeast.

(A) PolyQ-mediated Sup35 aggregation converts endogenous Sup35 from the soluble into the insoluble state. Soluble Sup35Q92 monomers are joined together by polyQ-mediated polar zipper formation leading to the assembly of insoluble β-sheet-rich protein aggregates. These structures then function as templates for the polymerization of endogenous Sup35 molecules. Sup35Q92-mediated Sup35 aggregation is triggered by the Q/N-rich PrDs present in both proteins. Sup35 aggregation is accompanied by conformational changes in the Q/N-rich PrD leading to the formation of aggregation-prone β-sheet-rich structures. (B) Template-mediated Sup35 aggregation. Pre-existing β-sheet-rich Rnq1 aggregates with an accessible Q/N-rich PrD stimulate the aggregation of soluble Sup35 molecules, which contain a related Q/N-rich PrD. Aggregation is accompanied by conformational changes in the Sup35 PrD leading to β-sheet-rich structures.