Research
The Bioprocess Laboratory engineers processes that employ biological catalysts for the production of biopharmaceuticals as well as pharmaceutical, fine and specialty chemicals. On the one hand, we integrate the ever-growing body of knowledge of the functioning of biological parts and systems into increasingly complex rational process designs. On the other hand, the still existing considerable gaps in our understanding of biological parts and systems force us to employ alternative design strategies, such as directed evolution, that can deal with uncertainty. The Bioprocess Laboratory explores the space that is spanned by these two vectors to contribute to the generation of novel bioprocesses, in particular in the areas of synthetic biology, alternative metabolisms, and separation-integrated processes.
We often apply the following methods or work in the following fields:
The Bioprocess Laboratory (BPL) has, over the last years, made major contributions to three different fields - integrated bioprocesses, synthetic biology and alternative metabolisms – employing rational forward design and naïve directed evolution to implement novel biocatalysts or novel process concepts.
On the one hand, working in vitro (with enzymes) allows rational design of even complex integrated reaction schemes, as we could show in our work of claiming isomerases (EC class V), one of the most underexplored enzyme classes in the realm of biotechnology, for efficient bioprocessing. Although there is plenty of chemical opportunity, the fact that isomerases consistently leave mixes of substrates and products has broadly prevented the industrial application of this class of enzymes. We provided what we consider a generic solution to this problem by model-based integration of a continuous high-performance separation method (simulated moving bed, a continuous chromatography process) with enzymes that had been engineered to optimally operate under the environmental conditions required for the integrated process to function effectively.
In this fashion, compound mixtures can be continuously separated in one step, and the remaining starting material returned to the enzyme to obtain a yield on starting material of 100%. We demonstrated the feasibility by implementing highly efficient process schemes in this fashion for the production of the rare sugar and low-calorie sweetener D-psicose from sucrose in a three-enzyme reaction, and also demonstrated principle feasibility for process schemes that are more demanding on the chromatography side, such as the formation of enantiopure fine chemicals.
Any attempt to transfer such a strictly model-based design approach to biosystems faces major challenges, including the fact that in vivo systems are not constant over short (adaptation to changes in the cell’s environment) and long (evolution) periods of time and that whatever property we look at results from the complex and non-linear interaction of multiple actors. Still, for process purposes it is essential to be able to optimize system performance, and therefore we undertook to apply forward design to a biosystem, though one of reduced complexity.
Specifically, we reconstituted a 10-step in vitro glycolysis and used it to produce starting materials for pharmaceutically relevant sugar derivatives. Next, we developed a real-time MS method that allows a highly accurate tracking of multiple compound concentrations in reactors, and then complemented this unparalleled analytical power with an experimental set-up that allowed application of a broad variety of complex perturbations on the system. Remarkably, these measures were sufficient to parameterize a mechanistic mathematical model to such a level of accuracy that we could forward design highly optimized production systems.
As mentioned, such levels of forward design are not (yet) possible with in vivo (cellular) systems. However, the advent of synthetic biology increasingly allows constraining the space that one has to sort through before a (close to) optimal biocatalyst solution can be identified. As a case in point, we again explored the question of optimal biosystem composition in terms of protein levels (as in the in vitro glycolysis above), but now not based on purified enzymes but on in vivo protein expression tailored via ribosome binding site (RBS) design. By developing algorithms that allow a drastic reduction of input RBS sequences for the combinatorial optimization of multi-protein catalytic systems, we could reduce an otherwise infeasible optimization problem to one that can be solved with a few hundred HPLC assays.
On the other end of engineering methodology are procedures that do not use prior information (or only relatively little) and go through the major effort of searching through a vast solution space (directed evolution). With the current advances in miniaturization and parallelization, also this approach has gathered much attraction, and we apply it to a broad variety of efforts ranging from improving vitamin producers to engineering novel antibiotics. More specifically, we use droplet-based microfluidics and sophisticated biosensing to solve questions that rely on the secretion of a specific product, such as a vitamin from an industrial production strain or a secreted antimicrobial peptide that is supposed to interact with a pathogenic target strain. We also construct the corresponding sensing modalities, ranging from RNA aptamers via transcriptional regulators to whole cell biosensors.
Finally, we apply these and also more traditional directed evolution methods to expand the biochemical sphere that is accessible to us, from enzymes that catalyze novel, new-to-nature reactions (such as metathesis) to novel peptides that might serve as last resort antibiotics in the fight against drug-resistant bacteria.
In summary, we try to employ the entire range of process engineering from rational design to naïve evolution procedures in order to find novel or improved biocatalysts and use them in innovative processes that give us access to novel and/or valuable compounds.