Filamentous Aggregates Are Fragmented by the Proteasome Holoenzyme

Abstract

Filamentous aggregates (fibrils) are regarded as the final stage in the assembly of amyloidogenic proteins and are formed in many neurodegenerative diseases. Accumulation of aggregates occurs as a result of an imbalance between their formation and removal. Here we use single-aggregate imaging to show that large fibrils assembled from full-length tau are substrates of the 26S proteasome holoenzyme, which fragments them into small aggregates. Interestingly, although degradation of monomeric tau is not inhibited by adenosine 5'-(3-thiotriphosphate) (ATPγS), fibril fragmentation is predominantly dependent on the ATPase activity of the proteasome. The proteasome holoenzyme also targets fibrils assembled from α-synuclein, suggesting that its fibril-fragmenting function may be a general mechanism. The fragmented species produced by the proteasome shows significant toxicity to human cell lines compared with intact fibrils. Together, our results indicate that the proteasome holoenzyme possesses a fragmentation function that disassembles large fibrils into smaller and more cytotoxic species.

Keywords: alpha-synuclein; disaggregation; proteasome; protein aggregation; tau; total-internal reflection fluorescence microscopy.

Graphical abstract

Figure 1

Imaging Fibrils with a Fluorescence TIRF Microscope (A) Recombinant full-length tau was aggregated for 24 h, and aliquots were taken and mixed with the proteasome in an ATP-containing proteasome buffer or with the buffer only as a control. After 0.5 h (starting reference) and 20 h of incubation, each reaction was diluted in an imaging buffer containing pFTAA. The chemical structure of pFTAA, which binds amyloid structures, is shown. (B) Samples were placed on a glass coverslip, excited with a 488 nm laser, and imaged on a custom-built TIRF microscope (see also Figure S3). A typical fibril (length, >1 μm) and diffraction-limited small aggregate (length, <1 μm) are shown. The scale bar represents 1 μm. (C and D) A large amount of fibrils remained present after incubation with the buffer alone (C), whereas treatment with the proteasome holoenzyme resulted in loss of fibrils and an increase in small aggregate count (D; depicted next to the 2D plot). The length of aggregates is plotted against the fluorescence intensity of pFTAA; the frequency is color-coded in the 2D plots. A processed image from each reaction is shown below the respective plots. The scale bars represent 10 μm. Results of three biological repeats (n = 3) performed independently using different protein preparations of tau and proteasome were combined into each plot. The SD between repeats was less than 20% in our TIRF experiments.

Figure 2

Fragmentation of Fibrils in the Absence of Soluble Tau Proteins (A and B) Aggregated tau samples were centrifuged. The pellet was resuspended in fresh proteasome buffer followed by incubation with (A) buffer control or (B) the proteasome holoenzyme for 20 h and subsequently imaged and presented as described in Figure 1. (C and D) Proteasome holoenzymes pre-treated with (C) Velcade (proteasome^Velcade) or (D) ATPγS (proteasome^ATPγS) were subsequently incubated with aggregated tau as in (B). (E) Instead of the holoenzyme, fibrils were also incubated with regulatory particles (RPs) and analyzed as above. Combined results of three independent experiments (n = 3) are shown.

Figure 3

Aggregation of Tau in the Presence of Proteasome Holoenzymes (A and B) Monomeric tau proteins at 2 μM final concentration were mixed with proteasome buffer and imaged after (A) 0.5 h or (B) 24 h, showing a substantial increase in both fibril and small aggregate levels. (C–E) Aggregation of tau in the presence of 40 nM final concentration of (C) untreated, (D) Velcade-treated, and (E) ATPγS-treated proteasome holoenzyme, measured after 24 h of incubation. Each plot contains the cumulative data from three independent measurements (n = 3).

Figure 4

Disordered Aggregates of Amorphous Structures Detected by TEM after Fragmentation by the Proteasome (A and B) Aggregated tau samples prepared as in Figure 2 were incubated with (A) an ATP-containing buffer or (B) proteasome holoenzyme for 20 h, stained with uranyl acetate, and imaged by TEM. Arrows indicate typical aggregated structures. (C and D) Aggregated samples incubated with (C) the buffer control or (D) the proteasome were immunolabeled with an anti-tau antibody, stained with uranyl acetate, and imaged by TEM. The scale bars represent 100 nm. Representative images are shown of at least three independent repeats.

Figure 5

Fragmented Species Trigger Cell Lysis (A) A lactate dehydrogenase (LDH) assay was used to detect lysis of HEK293 cells after incubation with either intact aggregated tau (column 1), the fragmented species (column 2), the buffer alone (column 3), or proteasome holoenzyme alone (column 4). Cells were tested for levels of lysis after incubation with fibrils treated with free RP (column 5) or with free RP alone as a control (column 6). (B) Aggregation buffers and tau fibrils are not toxic to cells. Monomers (column 1) or fibrils (column 2–4) at 1-, 10-, and 100-fold (10× and 100×) of the incubation concentration in (A) were incubated with the cells. Several controls including sodium chloride-sodium phosphate-EDTA (SSPE) buffer alone (column 5), SSPE buffer containing heparin (column 6), the pellet (containing tau fibrils, column 7), or the supernatant fraction (containing soluble aggregates and monomers, column 8) after centrifugation (STAR Methods) were also tested for cell lysis. As a positive control (column 9), fibrils were sonicated briefly and centrifuged to separate insoluble fibrils so that the supernatant could be added to cells at the same calculated concentration. Error bars represent SD of measurements from the mean of three independent experiments (n = 3).

Figure 6

Proteasome Fragments αS Fibrils into Cytotoxic Species (A and B) Aggregates were assembled from αS monomers for 24 h and subsequently incubated with (A) an ATP-containing buffer or (B) proteasome holoenzyme for 20 h and presented in 2D plots as in Figure 2. Combined data of three independent repeats are shown (n = 3). (C and D) The assay was repeated with untreated (C) or proteasome-treated (D) aggregated αS for TEM imaging. Arrows indicate typical aggregated structures. The scale bars represent 100 nm. (E) HEK293 cells incubated with untreated (column 1) and proteasome-treated (column 2) samples of aggregated αS proteins, the buffer (column 3), or the proteasome alone (column 4) were tested for cytotoxicity using the LDH assay as in Figure 5A. Mean values are shown, with error bars representing SD. All experiments were independently repeated at least three times (n = 3) using fresh αS and proteasome preparations.