Molecular basis for different substrate-binding sites and chaperone functions of the BRICHOS domain

Abstract

Proteins can misfold into fibrillar or amorphous aggregates and molecular chaperones act as crucial guardians against these undesirable processes. The BRICHOS chaperone domain, found in several otherwise unrelated proproteins that contain amyloidogenic regions, effectively inhibits amyloid formation and toxicity but can in some cases also prevent non-fibrillar, amorphous protein aggregation. Here, we elucidate the molecular basis behind the multifaceted chaperone activities of the BRICHOS domain from the Bri2 proprotein. High-confidence AlphaFold2 and RoseTTAFold predictions suggest that the intramolecular amyloidogenic region (Bri23) is part of the hydrophobic core of the proprotein, where it occupies the proposed amyloid binding site, explaining the markedly reduced ability of the proprotein to prevent an exogenous amyloidogenic peptide from aggregating. However, the BRICHOS-Bri23 complex maintains its ability to form large polydisperse oligomers that prevent amorphous protein aggregation. A cryo-EM-derived model of the Bri2 BRICHOS oligomer is compatible with surface-exposed hydrophobic motifs that get exposed and come together during oligomerization, explaining its effects against amorphous aggregation. These findings provide a molecular basis for the BRICHOS chaperone domain function, where distinct surfaces are employed against different forms of protein aggregation.

Keywords: BRICHOS; fibrillar and amorphous aggregation; molecular chaperone; oligomer model.

FIGURE 1

Bri2 BRICHOS interacts with client amyloidogenic peptides. (a) AlphaFold2 structure prediction of the human Bri2 BRICHOS domain. Face A of BRICHOS is located between the central β‐sheet (β1–7) and α‐helix 1 as indicated by the arrow. (b) AlphaFold2 prediction of human Bri2 BRICHOS with its endogenous amyloidogenic peptide Bri23. The secondary structure elements are labeled, with the β‐hairpin structure of Bri23 shown in orange. (c and d) SAXS analysis of Bri2 BRICHOS monomer with and without Bri23. Kratky (c) and Porod‐Debye (d) plots of monomeric rh Bri2 BRICHOS alone (red squares) and in the presence of a 10‐fold molar excess of Bri23 (orange circles). (e) MST measurements of rh Bri2 BRICHOS in the presence of different concentrations of Bri23 (the traces are shown in Figure S1g). Average normalized fluorescent signal (‰) with standard deviation is plotted against Bri23 concentration ranging from 150 to 0.07324 μM with logarithmic scale on the x‐axis. The dissociation constant K _D estimated from the MST curve is 9.5 ± 4.8 μM. The Bri2 BRICHOS was fused to an NT* solubility tag for all the experiments.

FIGURE 2

Endogenous and exogenous amyloidogenic peptides have the same binding site in Bri2 BRICHOS. (a) Two different constructs, that is, Bri2 BRICHOS‐Bri23 and Bri2 BRICHOS, derived from the Bri2 proprotein, were generated and expressed in E. coli. (b) ThT kinetics of Aβ42 (3 μM) aggregation in the presence of Bri2 BRICHOS‐Bri23 or rh Bri2 BRICHOS dimers (100% molar ratio to Aβ42). (c and d) The aggregation half time (𝜏_1/2) and maximum growing rate (r _max ) of Aβ42 with rh Bri2 BRICHOS‐Bri23 or rh Bri2 BRICHOS extracted from (Figure S4).

FIGURE 3

Global fitting of Aβ42 with rh BRICHOS. (a–c) Global fits (solid lines) of Aβ42 aggregation traces (dash lines) with different concentrations of rh Bri2 BRICHOS dimers, that is, 0%, 10%, 50%, and 100% (referred to monomeric Aβ42 molar concentration) with a secondary nucleation dominated model. The fitting was constrained such that only one rate constant, that is, k _n (primary nucleation), k ₂ (secondary nucleation) or k ₊ (elongation), was the free fitting parameter at each time, as indicated in each panel. (d–f) Global fits (solid lines) of the aggregation traces (dashed lines) with different concentrations of rh Bri2 BRICHOS‐Bri23 dimers were constrained in the same way as for Bri2 BRICHOS dimers, as indicated in each panel. (g–i) Global fits (solid lines) of Aβ42 aggregation traces (dashed lines) with different concentrations of wildtype rh proSP‐C BRICHOS (0%, 10%, 50%, and 100%) (referred to monomeric Aβ42 molar concentration) with a secondary nucleation dominated model. Again, the fitting was constrained to allow only one free rate constant at a time, as indicated in each panel.

FIGURE 4

Canonical chaperone activity of Bri2 BRICHOS‐Bri23 oligomers. (a and b) Characterization of Bri2 BRICHOS‐Bri23 oligomers. (a) Representative image of negative‐stain micrographs of rh Bri2 BRICHOS‐Bri23. (b) 2D classes of rh Bri2 BRICHOS‐Bri23 oligomers. Representative classes showing both 2‐fold (overlaid with red squares) and 3‐fold (overlaid with blue squares) symmetries. (c) Kinetics of aggregation of 0.6 μM citrate synthase (CS) at 45°C alone (black), in the presence of Bri2 BRICHOS‐Bri23 oligomers at 0.3 μM (magenta), 0.6 μM (green), 1.2 μM (blue) or 2.4 μM (purple). (d and e) Effects of rh Bri2 BRICHOS‐Bri23 or rh Bri2 BRICHOS oligomers on CS aggregation at different molar ratios (referred to monomeric subunits) of BRICHOS:CS. The x axis shows the relative molar concentration of BRICHOS to CS.

FIGURE 5

Cryo‐EM model of rh Bri2 BRICHOS oligomers. (a) cryo‐EM structure of the rh Bri2 BRICHOS oligomer. View along the 3‐fold axis of the reconstruction with D3 point group symmetry. The map before post‐processing. One segmented asymmetric unit is enclosed within the lines together with its adapted model. Scale bar 20 Å. (b) The post‐processed map following segmentation using the complete depicted biological assembly. (c and d) Top‐ and sideview of the D3 assembly of Bri2 BRICHOS with hydrophobic triplets Y65‐P67, Y70‐I72, M76‐V78 depicted with surface representation. (e–g) Top‐ (e) and sideview (f) of the BRICHOS oligomer highlighting the cysteine residues, C31 and C100. Two of the subunits, K (orange) and L (blue‐green), are shown separately in (g), only K to the left and the KL dimer to the right.