Two molecules can contain the same atoms yet behave differently because their three-dimensional arrangements are mirror images. Making the desired version is an important challenge in chemistry. A study involving EvoEnzyme explored whether computer-designed changes could alter which version an enzyme prefers to produce.
In research published in the Journal of the American Chemical Society in 2023, Patricia Gomez de Santos and colleagues redesigned fungal peroxygenases. Their results showed that a small collection of enzyme variants could provide substantially different reaction preferences.
How a peroxygenase adds oxygen
Peroxygenases use hydrogen peroxide to supply oxygen for reactions on organic molecules. Their catalytic machinery includes a heme group, an iron-containing cofactor held inside the protein.
Which part of a molecule receives oxygen, and from which orientation, depends partly on how the molecule fits near that catalytic centre. Changing the surrounding protein can therefore change the product, even when the underlying type of reaction stays the same.
The team began with PaDa-I, an already evolved version of a peroxygenase from the fungus Agrocybe aegerita. Collaborators included researchers affiliated with EvoEnzyme, CSIC, the Technical University of Munich, the Weizmann Institute of Science and Delft University of Technology.
Designing a small collection of alternatives
The researchers used FuncLib, a computational method combining evolutionary sequence information with Rosetta calculations. It proposes combinations of amino-acid changes that can reshape an enzyme’s active site while remaining compatible with a functional protein.
Of 30 designs tested, 24 were functionally expressed in yeast. The successful variants contained four or five changes in the catalytic core and were tested against a panel of model compounds.
Several variants altered the balance between mirror-image products. Others changed the position at which a substrate was oxygenated. The result was a collection of different catalytic options, rather than one enzyme that was best for every reaction.
What “tenfold improvement” actually means
For ethylbenzene, selected variants achieved catalytic efficiencies up to about ten times that of PaDa-I. This metric combines turnover with the substrate concentration needed to support activity; it should not be read as a tenfold increase in reaction speed under every condition.
In a separate styrene experiment, variant d4 achieved roughly 15 times the total turnover number of the parent. Total turnover counts product formed per enzyme molecule over the assay. It is a different measurement from catalytic efficiency.
Both comparisons were against the laboratory-evolved PaDa-I parent. Describing the results as a general improvement over “natural enzymes” would lose that important distinction.
From altered selectivity to practical synthesis
The value of this approach is the ability to explore different product preferences without testing an enormous random collection of variants. However, the reactions still have to be evaluated experimentally: a functional computer design does not automatically deliver the desired selectivity.
Industrial usefulness also depends on the intended substrate, catalyst lifetime, peroxide delivery, production cost and product purification. The study demonstrated useful enzyme-engineering results, rather than a proven reduction in factory emissions or the cost of manufacturing medicines.
Related questions about keeping an enzyme productive under demanding conditions appear in our report on the engineered PB92 protease.
Source
Gomez de Santos, P. et al. (2023). Repertoire of Computationally Designed Peroxygenases for Enantiodivergent C–H Oxyfunctionalization Reactions. Journal of the American Chemical Society, 145(6), 3443–3453. Read the study. Open full text.
Update note: The original date and URL are retained. This revision replaces promotional company language with a research-focused account and corrects the comparator for the reported performance gains.
