Ubiquitination of alpha-synuclein filaments by Nedd4 ligases

Abstract

Alpha-synuclein can form beta-sheet filaments, the accumulation of which plays a key role in the development of Parkinson's disease, dementia with Lewy bodies and multiple system atrophy. It has previously been shown that alpha-synuclein is a substrate for the HECT domain-containing ubiquitin ligase Nedd4, and is subject to ubiquitin-mediated endosomal degradation. We show here that alpha-synuclein filaments are much better substrates for ubiquitination in vitro than monomeric alpha-synuclein, and that this increased susceptibility cannot be mimicked by the mere clustering of monomers. Recognition by Nedd4 family enzymes is not through the conventional binding of PPxY-containing sequences to WW domains of the ligase, but it also involves C2 and HECT domains. The disease-causing alpha-synuclein mutant A53T is a much less efficient substrate for Nedd4 ligases than the wild-type protein. We suggest that preferential recognition, ubiquitination and degradation of beta-sheet-containing filaments may help to limit toxicity, and that A53T alpha-synuclein may be more toxic, at least in part because it avoids this fate.

Fig 1. Efficient in vitro ubiquitination of alpha-synuclein filaments by Nedd4.

Ubiquitination assays of Nedd4 with alpha-synuclein (antibody Syn-1). (A) Coomassie-stained gels of purified wild-type (WT) and A53T mutant alpha-synuclein. (B) Time course of polymerization of WT and mutant alpha-synuclein assayed by thioflavin dye fluorescence (see Methods). (C) Coomassie-stained gel of purified Nedd4 ligase. (D) In vitro ubiquitination assays using Nedd4 [200nM] with alpha-synuclein [200nM] showing efficient ubiquitination of filamentous (fil.) wild-type alpha-synuclein with strongly reduced modification of the filamentous A53T mutant and hardly any detectable ubiquitination of the monomeric forms (mo.). (E) Ubiquitination assays of Nedd4 [150nM] with alpha-synuclein filaments [600nM] in the presence of ubiquitin chain-type specific deubiquitinases as indicated [UbiCREST, [14]] demonstrating mainly K63, but also some K29 and K33, linkages.

Fig 2. In vitro ubiquitination of alpha-synuclein by different Nedd4-type ligases.

Western blots of in vitro ubiquitination assays using different members of the Nedd4 family [250nM] with alpha-synuclein [600nM] (antibody Syn-1). (A) Comparing the ability of the Nedd4-type E3 ligases to ubiquitinate monomeric (mo.) or filamentous (fil.) alpha-synuclein indicates a general preference for filaments. (B) Ubiquitination assays of different Nedd4 family members with filamentous alpha-synuclein with or without the ubiquitin chain-type specific deubiquitinase Trabid as indicated above the panel. In each case the ubiquitin chains were sensitive to Trabid, implying similar linkages.

Fig 3. Ligase-specific recognition of alpha-synuclein filaments.

(A) Cartoon of the modular architecture of Nedd4-type E3 ligases and overview of the constructs used in this study. (B) Western blots of in vitro ubiquitination assays comparing different Itch versions (full-length, ΔC2, WW mutants [175nM, left panel]) and full-length or HECT-only versions of Nedd4, Nedd4L and Smurf2 [100nM, right panel] in their ability to ubiquitinate alpha-synuclein [600nM, upper panels]. The lower panels (anti-ubiquitin blot) show overall ubiquitination, demonstrating the activity of the E3 ligase constructs by their ability to auto-ubiquitinate. Note that full-length Smurf2 is auto-inhibited in the presence of monomeric alpha-synuclein. (C) Activation of autoinhibited Smurf2 by alpha-synuclein filaments. (D) Western blots of in vitro ubiquitination assays comparing full-length, ΔC2 or HECT-only versions of Nedd4, Itch and Smurf2 [100nM] in their ability to ubiquitinate alpha-synuclein [400nM], indicating ligase-specific recognition of filaments (antibody Syn-1). Three sections of the same exposure of the same gel are shown.

Fig 4. Clustering of alpha-synuclein does not result in efficient ubiquitination.

(A) Schematic of the polymerization domains (DIX and TPR) fused to a soluble substrate, either the cytoplasmic domain of Ndfip2 containing three PY motifs or full-length alpha-synuclein with mutants that were used in this study. The location of the A53T mutation is indicated, but mutant protein was not used in Fig 4 (see Fig 5A). (B) In vitro ubiquitination assays using Smurf2 with purified recombinant proteins as indicated above the panels, showing efficient Smurf2 autoubiquitination when incubated with polymerization-competent DIX-TPR-Ndfip2 with intact PY motifs (D-T-N2), but not when all three PY motifs were mutated (D-T-PPAG). No Smurf2 autoubiquitination was observed with DIX-TPR-alpha-synuclein (D-T-Syn) and only modest effects were observed with an introduced weak PY motif (D-T-Syn PY1). (C) Western blots of His pull downs from HEK293T cells co-transfected with His-Ub, HA-tagged Nedd4, WWP2 and the GFP-tagged highly efficient polymerizing DIX-TPR fusions as indicated. N2-Pym, Ndfip2 with all three PY motifs mutated; N2-1xPY, Ndfip2 with one remaining LPxY motif. (D) Live-cell images showing single confocal sections of representative HeLa cells transfected with DIX-TPR-alpha-synuclein-EGFP either with polymerization defective M4 mutant (top left), wild-type alpha-synuclein (middle left) or introduced weak PY motif (bottom left panel) and co-transfected with mCherry-Nedd4 (center panels). Recruitment of Nedd4 into larger punctate aggregates was observed only in the presence of the weak PY motif (see arrows). Insets show enlarged versions of the outlined areas.

Fig 5. No difference in ubiquitination between clustered wild-type and A53T alpha-synuclein.

(A,C) Western blots of His pull downs from HEK293T cells co-transfected with His-Ub, HA-tagged catalytically dead (C/G) or wt Nedd4 and the GFP-tagged polymerizing DIX-TPR fusions (A) or Dvl2-synuclein fusions (C) and their mutant versions as indicated. Clustered constructs with artificially introduced weak PY motifs are more efficiently ubiquitinated. (B) Schematic of the Dvl2-alpha-synuclein fusions used in (C). (D) Live-cell images showing single confocal sections of representative HeLa cells transfected with Dvl2-alpha-synuclein-EGFP either with polymerization defective M4 mutant (top left), wild-type alpha-synuclein (middle left) or introduced weak PY motif (bottom left panel) and co-transfected with mCherry-Nedd4 (center panels). As observed with the DIX-TPR-synuclein fusions (see Fig 4D) recruitment of Nedd4 into the larger punctate aggregates was observed only in the presence of the weak PY motif (see arrows). Insets show enlarged versions of the outlined areas.