(218h) On the Systematic Integration of Metabolic and Processes Engineering: The Case of Kerosene-Producing Saccharomyces Cerevisiae

The so-called ‘omics’ revolution and the advent of synthetic biology have been hailed as breakthroughs for the understanding and engineering of biocatalysts. Especially in synthetic biology, it is assumed that the practice of the Design-Build-Test-Learn (DBTL) engineering cycles will engender new industry-ready bioprocesses. However, so far this has not been the case since nascent bioprocesses display a very high early stage failure rate. Therefore, the whole research and innovation process needs to be improved, ensuring that R&D pipelines are implemented to accelerate development, and produce industrially relevant processes. The microbial production of fuels and industrial chemicals has been identified as a promising alternative to address the depletion of fossil resources and the climate change, which is tightly correlated to anthropogenic activities. The development of efficient cell-factories requires systematic metabolic engineering of microbial strains to rewire the metabolic network towards the desired behavior.

A typical process towards a suitable strain for bio-production of fuels and chemicals follows the iterative DBTL cycle, commonly involved in engineering practices. Nowadays, this process is expensive and laborious; the improvement iterations may last up to eight years while the total cost to develop a commercial strain is on the order of 50 M$. To address this problem, the IBISBA European consortium is developing a platform designated to accelerate the end-to-end bioprocess development. The platform seamlessly provides services that apply to the different R&D steps involved in the bioprocess development, supporting in that way the microbial strain design, the fermentation and downstream process development, as well as the project knowledge asset management. The current work is tightly correlated with the IBISBA consortium efforts, aiming to connect the downstream process design and yielding separation costs, which are considered among the key viability determinants with the upstream microbial strain design procedure.

In this work, an efficient computational strain design workflow is proposed to identify metabolic interventions that succeed high product revenues while demanding minimum separation expenses. The systematic workflow comprises of five modules: In the first module, the Genome-scale Metabolic reconstruction (GEM) of a selected host organism is edited to include metabolic pathways towards a selected product portfolio and economic variables related to the upstream process and the potential product revenue. In the second module, a Mixed-Integer Linear Program (MILP) formulation is addressed to identify alternative sets of reaction eliminations that result in maximum revenue. In the third module, we sample the GEM allowed solution space that correspond to the alternative metabolic strategies and estimate the product stream composition. In the fourth module, based on the exit stream compositions we identify the optimal separation flowsheet and minimum cost for product recovery by solving the corresponding superstructure optimization problem. Finally, the average separation cost and product revenue are used to identify the most promising metabolic strategies.

As a case study, we applied our workflow to rationally design a kerosene producing S.cerevisiae strain for minimum downstream separation cost. To this direction, S.cerevisiae iMM904 GEM was adapted to include hydrocarbons’ producing heterologous pathways. The developed strain design framework was applied to create a pool of alternative metabolic strategies that yield in maximum revenue. Assuming aerobic cell culture conditions in a chemostat array with glucose as the sole carbon source, the models that correspond to the distinct strategies were sampled to estimate the exit stream composition and a distillation supertask problem was solved to identify the minimum separation cost. The applied methodology identified metabolic strategies up to 7-fold more efficient with respect to the initial strain.

The present formulation is the first to our knowledge that aims to bridge the strain design procedure with the downstream process synthesis, paving the way towards microbial strains tailor-made for sustainable biorefinery applications.