Fig.1

The b'xa' domain of PDI interacts with N-terminal domain of αsyn. (a) Model for PDI (red) primary sequence, PDI ab domain (pink) and PDI b'xa' domain (blue) are shown underneath. (b) Surface crystal structure of human PDI (PDB ID: 4EKZ). (c) Protein purities analyzed by SDS-PAGE. M, Marker; lane 1, αsyn1-60; lane 2, αsyn; lane 3, PDI ab; lane 4, PDI b'xa'; lane 5, PDI. (d) and (e) Spectral overlay of αsyn in the absence (black) and presence of equivalent amount of PDI (red) and PDI b'xa' (blue). Residues showing significant signal attenuation or shifts were marked. (f) Residue-resolved attenuation (I/I0) of αsyn in the presence of PDI (red), PDI ab (pink) and PDI b'xa' (blue). (g) Hydrophobic residues of αsyn predicted by OMH method, the binding and hydrophobic areas are colored gray

Fig.2

The effect on αsyn aggregation of PDI, PDI ab and PDI b'xa' detected by ThT experiment. (a) Aggregations kinetics of αsyn in the absence (black) and presence of equimolar PDI ab (pink), PDI b'xa' (blue) and PDI (red). (b) The fitted half-time (t1/2) values of aggregation and their statistical analysis. ns: not significant, ****: p < 0.0001; the dots represent the values obtained from four parallel experiments

Table 1

Aggregation kinetics data

Fig.3

NMR results showing the interaction between PDI b'xa' with the N-terminal domain of αsyn. Overlaid 1H-15N TROSY spectra of PDI b'xa' (red) with (blue) αsyn 1-19 (a), αsyn 30-42 (d) or αsyn 1-60 (g). E239 were selected as representative residues for global chemical shift perturbations (CSPs) during NMR titration. Residue-resolved CSPs of PDI b'xa' during titrations with αsyn 1-19(b), 30-42(e) and 1-60 (h). KD values were fitted from curves of residues with significant CSPs, as a function of αsyn 1-19(c), αsyn 30-42(f), and αsyn 1-60 (i) concentration. Residues showing significant CSPs were marked in spectra and mapped red on the crystal structure surface of PDI b'xa'. Threshold level was indicated as 'mean+standard deviation'. Disperse residues were verified and removed by the threshold of 'mean +2×standard deviation'

Fig.4

Hydrophobic binding mode of PDI with the N-terminal domain of αsyn. (a) Docking model of PDI binding αsyn1-40 (left panel) and a local zoom view of binding area (right panel). The relevant residues of αsyn1-40 involving in binding PDI were depicted as magenta sticks. HDOCK experiments were performed on the web server (http://hdock.phys.hust.edu.cn/) by uploading human PDI crystal structure, amino acid sequences of αsyn1-40 and their binding sites identified by NMR titration. Optimal docking results were plotted with PyMol software, PDI molecular surface was colored gray and its hydrophobic pocket was marked in red, the magenta cartoon represented αsyn1-40 structure. (b) Alignment of amino acid sequence of human PDI b'xa' from HiPDI b'xa'. Sequence identity and similarity were calculated by Clustalw[52] and ESPript 3.0[53]. Similar amino acids were showed as black letters in yellow boxes and identical amino acids were showed as white letters in red boxes. The secondary structural domains were indicated on the top of the sequences. Blue and green asterisks represented the key hydrophobic residues of αsyn interacting with human PDI and HiPDI, respectively