Computation, Synthesis and NMR Investigations of PROTAC Linker Conformation, Chamaleonicity and their Impacts on the Mode of Action

Linker design is a core task in Proteolysis Targeting Chimera (PROTAC) drug discovery, which influences ternary complex formation for selective target protein degradation as well as physicochemical properties. However, it remains challenging to accurately model the whole landscape towards the formation of ternary complex and understand the influence of linker rigidity/flexibility on each step in the working mechanism of PROTAC. To address this, we revisited a well-established system of the VHL-based BET degrader MZ1 and also explored the likely ternary complex of ARV771 with enhanced MD simulation, which explains for the first time the plasticity and selectivity of BET degradation with VHL E3-ligase. Regarding facile conformational change possibly induced by different linkers, the physics-driven modelling of ternary complex was applied to confirm and propose a highly cooperative and potent BET degrader HL1 which was shown to preferentially degrade BRD4 with a late onset of the hook-effect, similar to MZ1. A further novel PROTAC containing a semi-rigid fully sp3 linker, HL1CON, was developed to potentially stabilize the MZ1-type ternary complex similarly to HL1 and improve membrane permeability, but also appears to suffer from non-chameleonic hydrophobic collapse in different solutions as well as in VHL binary complexes. This resulted in significant barriers for HL1CON to get into cell, bind with either receptor and ultimately form cooperative and productive ternary complex. This study has demonstrated that rational PROTAC linker design is a multi-task object and the rigidification strategy with redundant hydrophobicity does not necessarily benefit degradation activity in our case. The comprehensive modelling of PROTACs in solution, binary complex and ternary complex states based on physics could help us to understand its complicated behaviors and reduce the risk in structure-based linker optimization.