Towards simulation-led engineering of a natural Diels-Alderase

Abstract

The Diels-Alder (DA) reaction is a fundamental reaction in organic synthesis, and the discovery of enzymes that can catalyse this reaction (or DAases) has thus long been a desirable goal. AbyU is a DAase which performs the cyclisation step in the formation of abyssomicin C, a spirotetronate antibiotic. There is therefore interest in engineering this enzyme to use it as a platform for more general spirotetronate synthesis. In this thesis, the origin of catalysis in AbyU is firstly studied via multiscale computational simulation. It is determined that the role of the active site is chiefly to provide complementary binding for the reactive conformation of the substrate. The potential for expanding the substrate scope of AbyU is then explored by predicting its activity (and that of several point mutants) with a panel of alternative substrates. A workflow of efficient computational protocols including docking, molecular dynamics (MD) and combined quantum mechanics / molecular mechanics (QM/MM) reaction simulation is therefore developed for screening the different combinations (utilising the same protocols which were used to study the native reaction). Due to the simplicity of the enzyme, it is found to accept most alternative substrates, which are all predicted to be fairly active with the enzyme. In addition, the enzyme was rationally redesigned for specific substrates, and redesigned variants were evaluated using the same in silico screening protocols. This demonstrates a promising approach to simulation-led engineering of this enzyme, where good scope for optimisation was suggested for one particular substrate. Overall, the in silico screening protocols developed in this thesis provide a valuable example of computationally efficient simulations that can be used to generate activity predictions for β-barrel spirotetronate forming enzymes such as AbyU, in a reasonably high throughput manor. The remaining aspects of enzymatic catalysis are then lastly studied. The origin of stereoselectivity is probed, where it is found that short MD simulations of the Michaelis complex alone are sufficient to indicate this. Finally, the behaviour of the catalytically important capping loop is studied and its potential for optimisation explored. A particular enhanced sampling technique shows utility for this, enabling the potentially rate limiting loop opening (for product release) step to be captured on practical simulation timescales

Date of Award	24 Jun 2021
Original language	English
Awarding Institution	University of Bristol
Sponsors	EPSRC RCUK
Supervisor	Marc W Van der Kamp (Supervisor) & Paul Race (Supervisor)