Insights into the stereoselectivity of human SETD7 methyltransferase

Abstract

Human SETD7 methyltransferase (hSETD7) is involved in a wide range of physiological processes, and has been considered as a significant target to develop new drugs. (R)-PFI-2, one hSETD7 inhibitor, could bind to the pocket of substrates with potent (low nanomolar) activity and high selectivity, while its enantiomer (S)-PFI-2 showed 500-fold less activity in IC₅₀ determination. Why do this pair of enantiomers, with nearly identical structures, exert tremendously different inhibitory activity? We performed a total of 900 ns long-timescale molecular dynamics (MD) simulations and 80 ps hybrid quantum mechanics/molecular mechanics (QM/MM) MD simulations to understand the molecular mechanism of the stereoselectivity of hSETD7. For each SAM/hSETD7/PFI-2 system, we characterized and compared the residual fluctuation of hSETD7, and generated molecular interaction fingerprints (IFP) to exemplify the propensities of SAM/hSETD7-inhibitor interactions. Based on the QM/MM MD, we accurately captured the difference of hydrogen bonds between the SAM/hSETD7/(R)-PFI-2 and SAM/hSETD7/(S)-PFI-2 systems. Especially the strength of the hydrogen bond between G336 and two inhibitors, which determines the stability of the post-SET loop. The energy barrier for (S)-PFI-2 was much bigger than (R)-PFI-2 from global minimum to bioactive conformation as the potential energy surface scanning (PES) showed. Moreover, by estimating the binding affinity and phylogenetic tree analysis, we discovered 16 hotspots were essential for binding both enantiomers but the specific mode of interaction between these hotspots and enantiomorphs is different. Our findings reveal the effect of chirality on the inhibition activity of hSETD7 in detail, and provide valuable information for hSETD7 structure-based drug development.

Fig. 1. RMSD and RMSF plots for (R/S)-PFI-2/SETD7/SAM systems. Black, red, green: trajectories of three independent MD simulations (run-1, run-2, run-3). (A) and (B) are the RMSD plots of (R)-PFI-2/SETD7/SAM system and (S)-PFI-2/SETD7/SAM system, respectively. (C) and (D) are the RMSF plots of (R)-PFI-2/SETD7/SAM system and (S)-PFI-2/SETD7/SAM system, respectively.

Fig. 2. The interaction fingerprints of the enantiomeric inhibitors with SETD7/SAM. (A) and (B) are the normalized frequency of hydrophobic interactions in the last 50 ns for (R)-PIF-2/hSETD7/SAM and (S)-PIF-2/hSETD7/SAM system, respectively. (C) and (D) are the normalized frequency of hydrogen bonds or ionic bonds in the last 50 ns for (R)-PIF-2/hSETD7/SAM and (S)-PIF-2/hSETD7/SAM system, respectively. Cyan, magenta, and bright-orange areas represent the three different MD simulations presented in this work.

Fig. 3. Electrostatic potential surfaces of QM region. The electrostatic potential is measured in eV, with range as shown in the corresponding color bar. Hydrogen bonds are displayed by yellow dashes. (A). (R)-PFI-2 and the key residues of binding site. (B). (S)-PFI-2 and the key residues of binding site.

Fig. 4. Potential energy surface (PES) obtained along DA1 and DA2 dihedral angles. B structures stand for the bioactive conformation. M structures stand for globe minimum.

Fig. 5. A phylogenetic tree of energy contribution for residues 117–363 of SETD7 and co-factor SAM.