We designed a novel pathway to a compound that has never before been produced in bacteria, using an innovative computational retrosynthesis method.
We selected promising enzymes that could carry out the predicted novel reactions using computational tools.
We used organic chemistry techniques to propose two spontaneous reactions which complete our desired pathway.
We used Flux Balance Analysis to predict the function of the novel pathway in the context of bacterial metabolism.
Even in a normal year modelling would have been a cornerstone of our project, but in a year where access to labs has been made impossible due to the Covid-19 pandemic modelling becomes the driving force for the proof of concept of a project.
We generated a workflow including computational retrosynthesis, pathway assembly and enzyme selection, coupled with manual chemical analysis and computational Flux Balance Analysis. All these techniques are further explained in our two main modelling pages, Pathway Design and Flux Balance Analysis.
Our aim this year was to engineer bacteria to produce a biosynthetic UV filter, more specifically Hipposudoric Acid, a compound naturally produced by Hippopotamus amphibius, the African hippo. There have not been previous attempts to produce Hipposudoric Acid in bacteria, and the biosynthetic pathway in the hippopotamus is not yet certain.
Our approach consisted of a mix of computational and non-computational methods, to come up with a novel pathway to produce our compound of interest in E. coli.
Our pathway discovery method is based on retrosynthesis, performed by the computational tool Retropath2.0, which finds sets of reactions from our target compound to possible precursors in E. coli’s metabolome.
The resulting reaction sets are processed with rp2paths, which validates them and organises them into potential biosynthetic pathways.
In a third step we predict enzymes that could catalyse the reactions in our novel pathways to Hipposudoric Acid, using the computation tool Selenzyme, which suggests enzyme candidates and ranks them according to their probability of catalysing our new target reactions.
To refine the results we used non-computational methods, in particular chemical analysis, which helped us to fill remaining gaps in the predicted pathways and validate the suggested sequence of reactions towards our target compound.
After defining our novel pathway towards Hipposudoric Acid, we decided to perform a Flux Balance Analysis (FBA), to understand how our pathway would operate in the context of bacterial metabolism. FBA is a method for analyzing the flux of metabolites through a metabolic network. Using the COBRA Toolbox in MATLAB, the main aim of FBA for our project was to predict optimal growth conditions for the production of our target compound.
To simplify the computational work we decided to do this analysis on a model of E. coli metabolism including the reactions in our novel pathway up to the reaction from 3-(4-hydroxyphenyl)pyruvate (34hpp) to Homogentisic acid (hga), the last precursor for Hipposudoric Acid which is produced in an enzyme-catalysed reaction. The insights gained from the FBA will help next year’s team on their experimental design to produce the compound in the laboratory, possibly by engineering the host strain metabolism.
From the structure of Hipposudoric Acid it can be easily seen that this molecule can be derived from two units of Homogentisic Acid. We proposed an oxidative dimerisation of HGA is the missing step for the production of Hipposudoric Acid. This oxidation might occur in two steps producing benzoquinone acetate, which is then further oxidised and dimerises to form Hipposudoric Acid. We evidence this using organic chemistry techniques and primary literature.
Acknowledgements
We want to thank Dr Ruth Stony and Dr Hong Zeng. All of our modelling efforts would not have been possible without the Support and Guidance from Ruth on the Pathway Design and Hong on the Flux Balance Analysis.
