Conformational Ensemble of hIAPP Dimer: Insight into the Molecular Mechanism by which a Green Tea Extract inhibits hIAPP Aggregation

Small oligomers formed early along human islet amyloid polypeptide (hIAPP) aggregation is responsible for the cell death in Type II diabetes. The epigallocatechin gallate (EGCG), a green tea extract, was found to inhibit hIAPP fibrillation. However, the inhibition mechanism and the conformational distribution of the smallest hIAPP oligomer – dimer are mostly unknown. Herein, we performed extensive replica exchange molecular dynamic simulations on hIAPP dimer with and without EGCG molecules. Extended hIAPP dimer conformations, with a collision cross section value similar to that observed by ion mobility-mass spectrometry, were observed in our simulations. Notably, these dimers adopt a three-stranded antiparallel β-sheet and contain the previously reported β-hairpin amyloidogenic precursor. We find that EGCG binding strongly blocks both the inter-peptide hydrophobic and aromatic-stacking interactions responsible for inter-peptide β-sheet formation and intra-peptide interaction crucial for β-hairpin formation, thus abolishes the three-stranded β-sheet structures and leads to the formation of coil-rich conformations. Hydrophobic, aromatic-stacking, cation-π and hydrogen-bonding interactions jointly contribute to the EGCG-induced conformational shift. This study provides, on atomic level, the conformational ensemble of hIAPP dimer and the molecular mechanism by which EGCG inhibits hIAPP aggregation.

The coordinate files for the dimers with and without EGCG will be available free of charge upon Email to ghwei@fudan.edu.cn.

Simulation details.
To avoid any bias of initial secondary structure on the simulation results, we heated the hIAPP monomer taken from protein data bank (PDB id: 2L86) at a high temperature of 750 K to generate extended coil states of hIAPP. Then we randomly selected 24 random coil conformations from the high-temperature MD-generated hIAPP monomers. We placed each two conformations in turn in a perpendicular orientation to avoid bias in favor of parallel or antiparallel β-sheet alignments during REMD simulations.
The minimum distance between the two hIAPP chains is 1.0 nm. Thus we obtained 12 initial states of hIAPP dimer for REMD simulations. The minimum distance from the hIAPP to the edge of water box is 0.8 nm. Counterions (Cl -) were added to neutralize 2 the two systems. The integration time step of REMD simulation is 2 fs. Each replica was run for 360 ns, giving an accumulated simulation time of 17.28 μs for each system.
The acceptance ratio for each REMD run is ~16% for both systems. Electrostatic interactions were calculated using the particle mesh Ewald method with a real space cutoff of 1.0 nm. The van der Waals interactions were calculated using a cutoff of 1.4 nm. The solute and solvent were separately coupled to external temperature and pressure baths.

Analysis details.
The two-dimensional (2D) free energy surface was constructed using -RT ln P(x, y), where P(x, y) was the probability of two selected reaction ordinates, x and y. The x and y coordinates used in this study include: 1) the radius of gyration (Rg) and the number of hydrogen bonds (H-bonds) (including intra-and inter-molecular H-bonds) of hIAPP dimer; 2) the centroid distance and the orientation between the aromatic rings of F15, F23, and Y37 residues and the three rings of EGCG molecules. An H-bond was considered as formed if the N…O distance is less than 0.35 nm and the N-H…O angle is greater than 150°. A contact was defined when the aliphatic carbon atoms of two nonsequential side-chains (or main-chains) come within 0.54 nm or any other atoms of two non-sequential side-chains (or main-chains) lie within 0.46 nm. The VMD program was used for graphical structure analysis.

Convergence check of REMD simulations.
The average coil/β-sheet probabilities over all residues within the two time intervals are 39.6/12.0% versus 40.3/10.6% for the hIAPP-dimer system and 47.8/3.5% versus 48.3/3.6% for the hIAPP-dimer+EGCG system ( Fig. S1 (A) and Fig. S2 (A)). The bend, turn and helix also have similar average probability within the two time intervals. The secondary structure contents for each residue within the two time intervals (Fig. S1 (B-F) for the hIAPP-dimer system and Fig. S2 (B-F) for the hIAPP-dimer+EGCG system) are also quite similar for both systems. The distributions of Rg and the total number of H-bond of the hIAPP dimer within the two time intervals also have much overlaps for both systems (Fig. S1 (H) and Fig. S2 (H)). The representative replica of the hIAPPdimer/hIAPP-dimer+EGCG system visited the full temperature space several times during the 360 ns simulation (Fig. S1 (I) and Fig. S2 (I)), demonstrating that the replica was not trapped in one single temperature. Taken together, these data demonstrate that the two REMD simulations are reasonably converged within 360 ns.  Figure S1. Simulation convergence assesments for hIAPP-dimer system using the data generated within 240-300 and 300-360 ns time intervals. We used the following several  and side-chain atoms of hIAPP dimer. Figure S6. Analysis of the cation-π interaction. The distribution of the minimum distance between the side chain NH3 + group of Arg and the center of each aromatic ring (A). A representative snapshot in which the distance between the side chain NH3 + group of Arg11 and the ring_1 of EGCG is 0.43 nm (B). This short distance indicates the existence of cation-π interaction. Figure S7. The chemical structure of an EGCG molecule. The red numbers label the carbon atoms in the three different aromatic rings: Ring-1, Ring-2 and Ring-3. Figure S8. The 12 initial states of hIAPP-dimer system. The peptides are shown by new-cartoon representation using VMD. Figure S9. The 12 initial states of hIAPP-dimer+EGCG system. The peptides are shown by new-cartoon representation using VMD.