Defective recognition of LC3B by mutant SQSTM1/p62 implicates impairment of autophagy as a pathogenic mechanism in ALS-FTLD

Abstract

Growing evidence implicates impairment of autophagy as a candidate pathogenic mechanism in the spectrum of neurodegenerative disorders which includes amyotrophic lateral sclerosis and frontotemporal lobar degeneration (ALS-FTLD). SQSTM1, which encodes the autophagy receptor SQSTM1/p62, is genetically associated with ALS-FTLD, although to date autophagy-relevant functional defects in disease-associated variants have not been described. A key protein-protein interaction in autophagy is the recognition of a lipid-anchored form of LC3 (LC3-II) within the phagophore membrane by SQSTM1, mediated through its LC3-interacting region (LIR), and notably some ALS-FTLD mutations map to this region. Here we show that although representing a conservative substitution and predicted to be benign, the ALS-associated L341V mutation of SQSTM1 is defective in recognition of LC3B. We place our observations on a firm quantitative footing by showing the L341V-mutant LIR is associated with a ∼3-fold reduction in LC3B binding affinity and using protein NMR we rationalize the structural basis for the effect. This functional deficit is realized in motor neuron-like cells, with the L341V mutant EGFP-mCherry-SQSTM1 less readily incorporated into acidic autophagic vesicles than the wild type. Our data supports a model in which the L341V mutation limits the critical step of SQSTM1 recruitment to the phagophore. The oligomeric nature of SQSTM1, which presents multiple LIRs to template growth of the phagophore, potentially gives rise to avidity effects which amplify the relatively modest impact of any single mutation on LC3B binding. Over the lifetime of a neuron, impaired autophagy could expose a vulnerability, which ultimately tips the balance from cell survival toward cell death.

Keywords: ALS; Atg8/LC3; FTLD; LIR; SQSTM1/p62; autophagy.

Figure 1.

ALS-FTLD-associated SQSTM1 mutations impact on the recognition of LC3B (SQSTM1^L341V) or ubiquitin (SQSTM1^G425R) in vitro. Mutations as indicated (or wild type, WT) were introduced into the full-length GST-SQSTM1 sequence and affinity isolation assays (LC3B and ubiquitin on beads) were performed at 37°C. Bacterial lysates containing the GST-SQSTM1 fusions were incubated with glutathione- (G), control- (C), LC3B (LC3), and ubiquitin-Sepharose (Ub) beads and captured proteins were detected by western blotting (anti-SQSTM1 antibodies). A representative blot is shown; see Fig. S1 for quantification of 3 independent experiments.

Figure 2.

ESI-MS indicates weaker binding of the LIR (L341V) to LC3B compared to WT LIR. (A) Native ESI-MS spectrum of an equimolar mixture of WT LIR and LIR (L341V) peptides (5 µM, residues 332 to 351). (B) LIR peptide mixture titrated with 5 µM LC3B. Top, full spectrum indicating free LIR (gray filled circle, mixture of WT LIR and LIR [L341V]), free LC3B and LC3B-LIR complexes of indicated (multiple) charge states. Below, zoomed-in spectra showing m/z values of the free and bound LIR complexes at the indicated charge states.

Figure 3.

Isothermal titration calorimetry (ITC) binding isotherms from titrating LC3B with LIR peptides at 25°C. Binding isotherms were fitted to a one-site binding model and K_d values determined. Starting concentration of the LC3B was approximately 30 µM, and of LIR peptide (residues 332 to 351) stocks approximately 350 µM. Black diamond, LIR (L341V); gray circle WT LIR.

Figure 4.

NMR titrations indicate selective perturbations of LC3B residues upon binding of LIR peptides. (A) ¹H-¹⁵N-HSQC spectrum of LC3B (0.25 mM dark gray) overlaid with the spectrum of the complex of LC3B with WT LIR (0.5 mM blue, residues 332 to 351) (ratio of 1:2) showing extensive chemical shift perturbations (CSPs) across the spectrum induced by ligand binding at 298 K; (B) Expansion of the region highlighted in (A) showing overlap between the spectrum of LC3B (dark gray), LC3B+WT LIR (blue) and LC3B + LIR (L341V) (green) illustrating a number of key residues within the LIR binding pocket that are perturbed to different extents by the WT LIR and LIR (L341V). Arrows identify the shifts of V33, F52 and V54 in the 2 complexes, while the majority of other residues show only small differences between the spectra of the 2 complexes (overlap of blue and green peaks), demonstrating selective effects on residues in direct contact with the LIRs.

Figure 5.

Chemical shift mapping of the binding of the WT LIR and LIR (L341V) peptides (residues 332 to 351) to ¹⁵N-LC3B. (A) Chemical shift mapping of the WT LIR. Weighted chemical shift pertubation (CSP) data showing the residues of ¹⁵N-LC3B (0.25 mM) that are perturbed upon WT LIR binding at a final ratio of 1:2 (LC3B:LIR) at 298 K. All CSPs above 1.0 are indicated. (B) Difference in CSP effects between WT LIR and LIR (L341V) sequences from NMR titrations at a final ratio of 1:2 (LC3B:LIR) at 298 K. The indicated residues showed CSPs that are substantially different between the 2 complexes (greater in the WT LIR). (C) Representation of binding surface of LC3B with WT LIR, generated by highlighting residues determined from NMR chemical shift mapping experiments. LC3B residues in blue showed the greatest CSP values. Backbone representation of the LIR peptide also indicated (residues DDDWTHLS, 335 to 342 shown). (D) Residues showing substantially different CSPs from the NMR titrations of LC3B with WT LIR compared to LIR (L341V), highlighted on the LC3B structure. Note the close correlation with the binding cleft in proximity to L341 of the LIR peptide (site of L341V indicated).

Figure 6.

Imaging of NSC-34 cells transfected with mCherry-EGFP-SQSTM1 constructs. mCherry-EGFP-SQSTM1 constructs (wild type [WT] or L341V mutant) were transiently transfected into NSC-34 cells. Thirty-two h post-transfection cells were left untreated or treated with BafA1, 20 nM and 16 h later visualized on a DeltaVision microscope. Images shown are representative of cellular phenotypes of the vast majority of cells in each treatment group. Scale bars: 10 µm. (RHS) intensity plots along the lines as indicated on the extended focus cell images showing co-incidence of EGFP and mCherry fluorescence.

Figure 7.

Quantification of colocalization of mCherry-EGFP-SQSTM1 transfected NSC-34 cells. Mean Pearson Correlation Coefficient (PCC) values of mCherry and EGFP overlap taken from a minimum of 50 cells per condition (NSC-34 cells transfected with mCherry-EGFP-SQSTM1 constructs) over a minimum of 3 independent experiments, with SEM indicated. Note the significant increase in PCC value in the wild-type-expressing cells treated with BafA1 and in the L341V mutant-expressing cells compared to wild type without BafA1 treatment (P values as indicated). When used cells were treated with 20 nM rapamycin 3 h before imaging (+Rap).