Electrostatically-guided inhibition of Curli amyloid nucleation by the CsgC-like family of chaperones

Abstract

Polypeptide aggregation into amyloid is linked with several debilitating human diseases. Despite the inherent risk of aggregation-induced cytotoxicity, bacteria control the export of amyloid-prone subunits and assemble adhesive amyloid fibres during biofilm formation. An Escherichia protein, CsgC potently inhibits amyloid formation of curli amyloid proteins. Here we unlock its mechanism of action, and show that CsgC strongly inhibits primary nucleation via electrostatically-guided molecular encounters, which expands the conformational distribution of disordered curli subunits. This delays the formation of higher order intermediates and maintains amyloidogenic subunits in a secretion-competent form. New structural insight also reveal that CsgC is part of diverse family of bacterial amyloid inhibitors. Curli assembly is therefore not only arrested in the periplasm, but the preservation of conformational flexibility also enables efficient secretion to the cell surface. Understanding how bacteria safely handle amyloidogenic polypeptides contribute towards efforts to control aggregation in disease-causing amyloids and amyloid-based biotechnological applications.

Figure 1. Phylogenetic distribution of CsgC- and CsgH-like sequences.

Distinct families of CsgC- and CsgH-like sequences occur across gram-negative bacterial phyla. The CsgC group is tightly-defined whereas CsgH can be split into several sub-types. The red lines connecting dots represent relatively-close sequence homology.

Figure 2. CsgH is a structural and functional homologue of CsgC.

(A) Ensemble of 20 lowest-energy structures calculated for CsgH (PDB 2N59). (B) Cartoon structure of the lowest-energy CsgH structure coloured according to the rainbow from N- (blue) to C-terminus (red). (C) Crystal structure of CsgC (PDB 2Y2Y) shown for comparison. Addition of (D) CsgH or (E) CsgC, respectively, at a range of substoichiometric molar ratios leads to a dose-dependent inhibition in the kinetic profile of CsgA amyloid formation within the ThT assay. (F) Addition of CsgC (solid lines) or CsgH (dashed lines) at various substoichiometric ratios results in a dose-dependent inhibition of FapC amyloid formation.

Figure 3. Negative stain EM images documenting a time-course for nucleating fibres.

Representative images from (A) CsgA and (B) CsgA + CsgC fibre formation are shown. CsgC was added at a substoichiometric ratio of 1:200. Time points for 0, 0.5, 2, 5 and 22 hours are shown (see additional. Scale bar = 200 nm. Additional examples for each time point are shown in Figs S2 and S4. (C) 2D Class average and corresponding raw images for CsgA double filament architecture. Scale bar, 16 nm. Panels (D,E) show ATR-FTIR and CD spectra, respectively, of CsgA fibres grown for 24 hrs in the absence or presence of CsgC (200:1 molar ratio).

Figure 4. Kinetic analysis of the effect of CsgC on CsgA amyloid formation.

(A) Addition of CsgC at a range of substoichiometric ratios results in progressive inhibition of CsgA, which was fitted to a nucleation-elongation model [cite Nat Prot]. The change in k + kn with inhibitor concentation is shown inset. (B) Addition of CsgC 1:400 at timepoints indicated by legend (in hours). The series truncated at these timepoints can be treated as seeded assays and fitted to a seeded nucleation-elongation model. This allows for the relative contributions of inhibition of nucleation and of elongation to the overall inhibition of k + kn to be separated. This clearly demonstrates a primary inhibitory effect on nucleation with a smaller effect on elongation.

Figure 5. Loss of charged residues in CsgC or CsgH reduces inhibitory potency.

Panels (A,B) show the relative potency of site-directed mutants of CsgC or CsgH, respectively within the ThT assay. (Define Quad and triple mutants). The raw fluorescence data were smoothed by a Savitsky-Golay filter for clarity. The molar ratio of CsgC/H to CsgA was 1:200. The ‘Triple’ CsgH mutant was T16D/Q18S/V20S. The ‘Quad’ CsgH mutant was K32A/D36S/R45S/K47E. The structural context of each mutation is shown for CsgC and CsgH in panels (C,D), respectively. Panels (E,F) show the electrostatic surface potential of CsgC and CsgH, respectively. The top sub-panels in each show a conserved positively-charged surface in both proteins.

Figure 6. Electrostatic screening effects of NaCl on CsgA amyloid formation and inhibition by CsgC.

(A) Addition of 0–500 mM NaCl to CsgA samples in the ThT assay causes a minor, dose-dependent reduction in the rate of amyloid formation. The intensity of ThT fluorescence also decreased by up to ~60% across the range Fig. S12. (B) The inhibitory potency of CsgC is gradually reduced as the NaCl concentration is increased. CsgC was added at a substoichiometric molar ratio of 1:200 throughout.

Figure 7. General model that describes inhibition of CsgA by CsgC.

The top panel indicates the general folding pathway for CsgA from disordered states via a key nucleation-competent intermediate (CsgA*) that productively multimerises and matures into amyloid fibrils. Although drawn as a monomer, CsgA* may be an oligomer. Fibril ends also capture appropriately structured monomers. Our data are consistent with a model whereby CsgC interacts with disordered monomers or CsgA* to prevent nucleation, effectively increasing the activation energy (E_act) required to form CsgA*. Within the cell (lower panel) this interaction prevents CsgA* formation and keeps CsgA in state readily picked up by CsgE for secretion through CsgG.