Phosphorylation of tau at a single residue inhibits binding to the E3 ubiquitin ligase, CHIP

Abstract

Microtubule-associated protein tau (MAPT/tau) accumulates in a family of neurodegenerative diseases, including Alzheimer's disease (AD). In disease, tau is aberrantly modified by post-translational modifications (PTMs), including hyper-phosphorylation. However, it is often unclear which of these PTMs contribute to tau's accumulation or what mechanisms might be involved. To explore these questions, we focus on a cleaved proteoform of tau (tauC3), which selectively accumulates in AD and was recently shown to be degraded by its direct binding to the E3 ubiquitin ligase, CHIP. Here, we find that phosphorylation of tauC3 at a single residue, pS416, is sufficient to weaken its interaction with CHIP. A co-crystal structure of CHIP bound to the C-terminus of tauC3 reveals the mechanism of this clash, allowing design of a mutation (CHIP^D134A) that partially restores binding and turnover of pS416 tauC3. We confirm that, in our models, pS416 is produced by the known AD-associated kinase, MARK2/Par-1b, providing a potential link to disease. In further support of this idea, an antibody against pS416 co-localizes with tauC3 in degenerative neurons within the hippocampus of AD patients. Together, these studies suggest a molecular mechanism for how phosphorylation at a discrete site contributes to accumulation of a tau proteoform.

Fig. 1. Phosphorylation of tauC3 inhibits interaction with CHIP.

A Cartoon schematic of the live-cell NanoBiT assay for measuring CHIP-tau PPIs (NanoBiT PDB = 7SNX; CHIP PDB = 2C2L). B Results of NanoBiT assays for CHIP interactions with tau proteoforms following treatment with okadaic acid (30 nM, 18 h) or DMSO control. Luminescence normalized to CHIP binding to FL 0N4R tau treated with vehicle (dashed line). Error bars represent standard deviation (SD). Statistical significance was determined by two-way ANOVA with Bonferroni’s post hoc analysis (n = 3 biological replicates). CTRL = mock transfected control. C Tau proteoforms binding to immobilized CHIP measured by ELISA. Assay was performed in technical triplicate and normalized to maximum absorbance at 450 nm. D Dissociation constants derived from (C). Error bars represent SD. E In vitro ubiquitination of tau proteoforms by CHIP. Samples were collected at the denoted timepoints, quenched in SDS-PAGE loading buffer, and analyzed by western blot. Assay was performed once.

Fig. 2. Phosphorylation of tauC3 Ser416 is sufficient to inhibit interaction with CHIP.

A Cartoon depicting constructs used for mapping of inhibitory phosphorylation sites by NanoBiT live cell assays. B NanoBiT live cell assay for CHIP interactions with truncated tau constructs following treatment with okadaic acid (30 nM, 18 h) or DMSO control. Data are shown as the difference between okadaic acid treated samples and vehicle control. Error bars represent SD. Statistical significance was determined by two-sided, one-way ANOVA with Tukey’s post hoc analysis (n = 3 biological replicates). C Apparent change in melting temperatures (Tm_app) for CHIP incubated with DMSO control or Hsp70 C-terminal peptides, as determined by DSF. Error bars represent SD. (n = 4 technical replicates). D DSF experiments performed and analyzed as in the previous panel, but with tauC3-derived peptides. (n = 4 technical replicates). E Competition FP experiment showing displacement of fluorescent tracer from the CHIP TPR domain by various 10-mer Hsp70 or tauC3-derived peptides. Experiments were performed in technical quadruplicate. F Inhibition constants derived from (E). Error bars represent SD. G Tau proteoforms binding to immobilized CHIP, as measured by ELISA. Assay was performed in technical triplicate and normalized to maximum absorbance at 450 nm. H Dissociation constants derived from (G). Error bars represent SD.

Fig. 3. TauC3 Ser416 phosphorylation regulates CHIP-dependent tau homeostasis.

A In vitro ubiquitination of tau proteoforms by CHIP. Samples were collected at the denoted timepoints, quenched in SDS-PAGE loading buffer, and analyzed by western blot. B Quantification of unmodified tau remaining following in vitro ubiquitination of varying tau proteoforms by CHIP. Unmodified tau remaining was analyzed by densitometry, normalized to 0-min time point, and curves were fit using one-phase exponential decay (n = 3 technical replicates). C Cartoon schematic depicting promoter architecture and varying C-terminal sequences for HEK293 FlpIn T-Rex cells expressing doxycycline inducible GFP-tau proteoforms. D Representative fluorescence micrographs for GFP-tau cells. Images show tau species associated with microtubules (green, false color), while nuclei are stained with Hoechst 33342 (blue, false color). Scale bar = 30 µm. Images are representative of three separate acquisitions. E Co-immunoprecipitation assay following IP of varying tau species from cells. Whole cell lysate (input) was used for loading controls, and co-immunoprecipitated CHIP was analyzed by western blot. Experiment was repeated in duplicate. F Representative western blot showing differing abundance of various tau proteoforms in HEK293 FlpIn T-Rex cells. Experiment was repeated in duplicate. G Quantification of tau protein abundance taken from three independent biological experiments. Tau:tubulin ratio was determined by densitometry and normalized to full-length tau. Error bars represent SD. Statistical significance was determined by one-way ANOVA with Tukey’s post hoc analysis. H Quantification of MAPT mRNA from three independent biological experiments. MAPT mRNA was normalized to GAPDH and shown relative to full-length tau. Error bars represent SD. Statistical significance was determined by one-way ANOVA with Tukey’s post hoc analysis.

Fig. 4. Structural basis for CHIP binding to tauC3 and inhibition by phosphorylation.

A 1.8 Å crystal structure of the CHIP TPR domain bound to a 10-mer tauC3 C-terminal peptide. The CHIP TPR domain is depicted in gray, while the tauC3 peptide is in yellow. Helices 1-7 of the CHIP TPR domain are numbered H1-H7. B Close-up view of interactions of tauC3 D421 and C-terminus with CHIP carboxylate-clamp residues K95, N65, N35, and K30. C Close-up view of interaction of tauC3 D418 with CHIP carboxylate-clamp residue K72. D Close-up view of interactions of tauC3 S416 with CHIP carboxylate-clamp residue D134 and proximity to CHIP F131. E Apparent melting temperatures (Tm_app) of CHIP WT or mutants in the absence or presence of 10-mer tau peptides as derived from DSF experiments. Error bars represent SD. (n = 4 technical replicates). F Competition FP experiment showing displacement of fluorescent tracer from WT or D134A CHIP TPR domain by tauC3 pS416 10-mer peptide. Samples were performed in technical quadruplicate and normalized to DMSO control. G Inhibition constants for tauC3 pS416 peptide for WT or D134A CHIP as derived from (F). Error bars represent SD. H In vitro ubiquitination of tau proteoforms by WT or D134A CHIP. Samples were collected at the denoted timepoints, quenched in SDS-PAGE loading buffer, and analyzed by western blot.

Fig. 5. TauC3 co-accumulates with Serine 416 phosphorylation in Alzheimer’s disease patient brains.

A Representative micrographs of immunofluorescence staining of tau pS416 (green), tauC3 (red), and nuclei (Hoechst 33342, blue) from the hippocampal CA1/CA2 region of human patient brains across increasing spectrum of Alzheimer’s Disease Neuropathological Change (ADNC) scoring. Scale bars = 50 µM; insets = 10 µM. B Quantification of tau pS416 staining across ADNC score. Box plot bounds dictate 25th−75th percentile, whiskers define sample minima and maxima, and line shows sample median. Analysis was performed on four unique patient samples for each score. Statistical significance was determined by one-way ANOVA with Tukey’s post hoc analysis. C Quantification of tauC3 staining across ADNC score. Box plot bounds dictate 25th−75th percentile, whiskers define sample minima and maxima, and line shows sample median. Analysis was performed on four unique patient samples for each score. Statistical significance was determined by one-way ANOVA with Tukey’s post hoc analysis. D Two-sided Pearson’s analysis showing correlation of increasing tau pS416 and tauC3 across ADNC score.

Fig. 6. MARK2 inhibits CHIP-dependent ubiquitination of tauC3, in part, by phosphorylating serine 416.

A In vitro phosphorylation of varying tau proteoforms by MARK2. Varying tau species were incubated overnight with recombinant MARK2, and relative phosphorylation of Ser416 was visualized by western blot. Ponceau S was used as loading control. Image is representative of duplicate experiments. B In vitro ubiquitination of unmodified tauC3 or MARK2-phosphorylated tauC3 by CHIP. Samples were collected at the denoted timepoints, quenched in SDS-PAGE loading buffer, and analyzed by western blot. C Quantification of unmodified tau remaining in (B). Samples were normalized to 0-min time point and curves were fit using one-phase exponential decay (n = 3 technical replicates). D Cartoon depicting the role of tau PTMs on interactions with CHIP and shuttling to various degradation pathways.