Research What we do
Evolution is an all-purpose problem solver. In nature, recursive cycles of diversification, selection, and amplification gave rise to a plethora of biomolecules with remarkable functions. Enzyme engineers, who mimic the Darwinian algorithm in the laboratory, aim to emulate natural selection and create biocatalysts with tailored activities. But classic directed evolution campaigns (Panel A) are notoriously labor- and time-intensive, thus limiting evolutionary searches to just a few cycles along a few trajectories.
Enter continuous evolution (CE), an umbrella term for approaches that achieve cycles of rapid mutation, selection and amplification inside living cells (Panel B).4 By minimizing human intervention and enabling parallelization, CE approaches can autonomously explore longer and more numerous evolutionary trajectories (Fig. 1C). However, current CE strategies are largely restricted to engineering enzymes of little value for industrial applications.
In our lab, we are developing a scalable and versatile CE-platform to autonomously engineer industrially-useful biocatalysts at an unprecedented rate and scale. Performing CE along long and many evolutionary trajectories will elucidate the sequence-structure-function relationships of these biocatalysts and enable us to train and improve machine-learning algorithms for the on-demand prediction of efficient enzymes, whose potential as sustainable alternatives to chemical syntheses has remained unexploited.
Macrocyclic peptides are privileged structures in drug discovery efforts. They ideally combine the beneficial properties of biologics (i.e. antibodies) and small-molecule drugs and thus, can engage with targets that are ‘undruggable’ by small-molecule therapeutics (e.g. protein-protein interfaces). Cyclic peptides are also superior to their linear counterparts, displaying higher binding affinities, greater stabilities, and improved membrane permeability. Given these advantages, the market for peptide pharmaceuticals – two thirds of the FDA and EMA-approved therapeutics are cyclic peptides – is expected to grow from 25 billion USD in 2018 to 50 billion in 2027.
The discovery of macrocyclic peptides that enter the drug pipeline as lead compounds is greatly accelerated by the use of biological selection platforms, which allow for the facile production and evaluation of vast (cyclic) peptide libraries (108-1013 library members). However, lead compounds elicited by such efforts typically require time-consuming and expensive chemical optimization efforts, which largely focus on their diversification with synthetic moieties to tailor pharmacological properties and binding affinities. As a result, strategies that are compatible with biological selections and can be employed to both cyclize and further diversify peptides are highly sought after as the resulting hybrid macrocyclic peptides provide better starting points for lead optimization (see Scheme above).
To maximize applicability in cyclic peptide discovery campaigns in academia and industry, we develop cyclization strategies that (1) proceed rapidly and selectively under mild conditions (i.e. in aqueous buffers at (near) neutral pH) and be fully compatible with routinely-employed biological selection platforms; (2) possess a great diversification potential to enable the incorporation of synthetic moieties akin to those found in cyclic peptides sourced from nature or those routinely employed by medicinal chemists.(3) come at the cost of a minimal chemical scar (=leftover functional groups, handles, or undesired moieties) upon successful peptide cyclization.