Engineering Specificity into Unspecific Peroxygenases (UPOs)

Engineering Specificity into Versatile Peroxygenases for Industrial Biocatalysis

Peroxygenases are a fascinating class of enzymes that are gaining a lot of traction in biocatalysis due to their unique and versatile catalytic capabilities. They are capable of performing a wide array of oxyfunctionalizations of non-activated carbons.

This enzyme class is mechanistically simpler than P450BM3 as it requires only hydrogen peroxide as the terminal oxidant without the need for cofactors and appropriate cofactor recycling systems. Making them easier to use in an industrial manufacturing setting.

However, unspecific peroxygenases (UPOs) have been shown to exhibit remarkably high substrate promiscuity. Our goal here was to begin with a wildtype (WT) enzyme that could catalyze many different reactions and accept many different substrates, then use ultrahigh throughput screening (uHTS) to engineer specificity into the biocatalyst (Fig. 1).

Figure 1. Engineering specificity into a promiscuous biocatalyst.

Building a Diverse Library of AbrUPO Variants in Pichia pastoris

For this work, we built a diverse library of AbrUPO variants, previously shown to catalyze production of at least 51 products from 16 different substrates.¹ We expressed this library in Pichia pastoris, chosen because it can perform glycosylation and it is known to secrete heterologous proteins. The theoretical diversity of this library was 5 x 10⁶ variants, and long-read sequencing confirmed at least 10⁴-10⁵ variants in the library (Fig. 2).

Figure 2. Long-read sequencing identified UPO variants in different fluorescently sorted fractions.

Screening UPO Activity Using a Fluorimetric Microfluidics Assay

We developed a fluorimetric assay compatible with our microfluidics platform that allowed us to screen this library for variants that were active against either a desired substrate, styrene, or an undesired substrate, propylbenzene (Fig. 3).

Figure 3. a) Schematic showing experimental workflow. Pichia cells containing genomically integrated UPO variants are singly encapsulated in pL-sized fluorous oil emulsion droplets along with media and an inducing agent. Cells are allowed to multiply and express UPO enzymes within each droplet. Substrate is then injected into each droplet, along with H2O2 necessary for UPO catalysis. A sensor is then injected into each droplet that converts residual H2O2 into a fluorescent species. Droplets are then collected in two separate channels, with bright droplets enriched for inactive enzymes or empty droplets and dark droplets enriched for active enzymes. b) Substrates and potential products generated by UPO enzymes. c) Images of microfluidic droplets as they progress through the workflow.

Validating UPO Variants with GC and HPLC Analysis

Of the tested lead compounds, UPO 5 is more active than WT, producing 3-5x more of each product, and UPO 20 is more specific than WT, producing 78% as much styrene oxide, 33% as much phenylacetaldehyde, and 24% as much phenylpropanol as WT. (Fig. 4).

Figure 4. Styrene products (top) were characterized by GC and propylbenzene products (bottom)were characterized by HPLC.

Conclusion

In a single round of directed evolution, we screened a diverse library containing at least 10⁴ AbrUPO variants expressed in Pichia pastoris. Using a modular microfluidics assay, we engineered one AbrUPO variant that produces purer styrene oxide product and another variant that catalyzes 3-5x more production than the parent enzyme.

¹Schmitz, F. et. al., React. Chem. Eng., 2023.