Submission Details

Design Rationale:

The methylsulphonyl group of fragment x72 sits in a pocket with multiple hydrogen bonding possibilities e.g. sidechains of T25, H41 and N142 so switching from methylsulphonyl to sulphonamide gives various conformational possibilities predicted to enhance affinity. The the aromatic group also co-locates with aromatics and other hydrophobics of several other fragments including x967. The first carbon up from the aromatic of x72 is close to one bond-length away from the N of the amide connector for the x967 hydrophobic group and comparison with other ligands shows some play in the aromatic orientation. I therefore started by connecting x72 to x967 with its hydrophobic group removed. Fragment 967 is potentially very useful as it has very good 3D-shape complementarity, and the phenol has induced a sidechain flip in N142 creating parallel surface and nicely burying the phenol aromatic group. Phenol oxygen pincered between His163 and main-chain oxygen of Phe140 with 2 H-bonds. Fragment 967 is also well positioned for extension into the cleft between Pro168 and E189, which region is largely unexploited in the initial fragment set. Terminal carbon in this direction ~3.3A from main-chain oxygen of E166, potential enhancement by shift from amide to ureido with further extension into the P168/E189 cleft.

Other Notes:

Compound design / selection cycle: Converted X967 compound to ureido and cut to amide at the other end then searched Chemspace REAL database for fragments so that any suggested ligands identified would be chemically reasonable. ~1000 fragments downloaded from Chemspace and linked to x72 fragment using Datawarrior. Several different chemophore pattern searches run with Pharmit server with chemophores from both x72 and x967 included in the alignment set. Higher value given to low RMSD shift than to calculated score. ~15,000 ligand conformations assessed by eye and favourite high scoring set downloaded and combined with the x967 protein model. Upload only supports one pdb so all ligand models added to this - ensemble set is the Autodock Vina energy minimisation output from Pharmit i.e. rigid-body protein and flexible ligand docking. Models generally give scores of 8-9 in CSM-lig. As a final point the ligands do contain 2 chiral centres, but they are designed to be made in sequential steps and you can already get single enantiomer versions of the component parts. I'm not a chemist, but I think the easiest route would be to have the middle bit with protected carboxy terminus, add on the ureido adducts and then deprotect / react with the other part (stereospecific version with amine available from Enamine).

Inspired By:

AAR-POS-d2a4d1df-16

AAR-POS-d2a4d1df-1