Lignin, a complex polymer in lignocellulosic biomass, is an abundant and renewable source of aromatic compounds. Reductive catalytic fractionation (RCF) efficiently produces phenolic monomers from lignin, including 4-n-propylguaiacol, a compound not currently used industrially. Isoeugenol, a valuable flavor and fragrance compound with applications in various industries (deodorants, vanillin synthesis, fine chemicals, polymers, and epoxy resins), is structurally similar to 4-n-propylguaiacol, differing only by an alkene double bond. A one-step selective dehydrogenation of 4-n-propylguaiacol to isoeugenol would represent a significant advancement in lignin valorization. Various enzymes, particularly the vanillyl alcohol oxidase (VAO) family, act on aromatic compounds. However, VAO typically produces a mixture of hydroxylated and dehydrogenated products. Bacterial eugenol oxidase (EUGO) presents a more attractive alternative due to its high expression levels in *E. coli*, homodimeric structure (simplifying computational analyses), solvent tolerance, and structural information available for enzyme engineering. Despite its homology to VAO, EUGO exhibits weak activity on 4-alkylphenols. This work focuses on engineering EUGO to efficiently and selectively catalyze the conversion of 4-n-propylguaiacol to isoeugenol.

Extensive research has explored the microbial catabolism of aromatic monomers and polymers, leading to the discovery of numerous enzymes capable of acting on aromatic compounds. Among these, the flavin-dependent VAO subfamily, including VAO, EUGO, 4-ethylphenol oxidase (EPO), eugenol hydroxylase (EH), and p-cresol methylhydroxylase (PCMH), stands out for its action on 4-alkylphenols. While structurally related and containing a covalently bound FAD cofactor, these enzymes exhibit diverse catalytic activities, including alcohol oxidation, hydroxylation, amine oxidation, oxidative demethylation, and dehydrogenation. VAO from *Penicillium simplicissimum* has been extensively studied, revealing its mechanism involving hydride transfer from the substrate to the flavin cofactor and subsequent formation of a p-quinone methide intermediate. This intermediate can undergo either hydroxylation (by water attack) or tautomerization to the dehydrogenation product. The reaction outcome is determined by the relative rates of these competing reactions, influenced by water accessibility and substrate orientation. VAO, however, exhibits a strong preference for hydroxylation over dehydrogenation and suboptimal expression in *E. coli*. EUGO from *Rhodococcus jostii* RHA1, a bacterial homolog of VAO, offers better expression and solvent tolerance but lacks sufficient activity towards 4-alkylphenols. This motivated the engineering of EUGO for the specific conversion of 4-n-propylguaiacol to isoeugenol.

The study employed a multi-step enzyme engineering approach, combining computational predictions with experimental validation and structural analysis. First, the researchers used FRESCO (Framework for Rapid Enzyme Stabilization by Computational libraries), a computational method combining FoldX, Rosetta calculations, and molecular dynamics simulations, to predict mutations that enhance thermostability. A library of 72 single mutants was screened for thermostability via thermal unfolding analysis (Thermofluor), identifying five mutations that significantly improved thermostability without compromising activity on vanillyl alcohol. These mutations were iteratively combined, resulting in a fivefold mutant (EUGO5X) with a 13.5 °C increase in Tm. Next, to enhance chemoselectivity towards dehydrogenation, the researchers used Autodock VINA and Rosetta Coupled Moves to identify active site mutations that improve protein-ligand interactions. A small library of mutants was screened, revealing that the S394A and S394V mutations strongly favored isoeugenol formation. The S394V mutation was introduced into EUGO5X. X-ray crystallography of the resulting S394V-EUGO5X mutant revealed that the improved chemoselectivity is due to reduced water accessibility to the active site. Despite improved selectivity, S394V-EUGO5X exhibited low activity. Further biochemical studies and structural analysis showed that a slow-decaying covalent substrate-flavin adduct was responsible for this low activity. Therefore, structure-guided mutagenesis was performed to target residues (D151 and Q425) near the flavin N5, aiming to destabilize this adduct. This led to the identification of two additional mutations (D151E and Q425S), resulting in the final eightfold mutant, PROGO. The crystal structure of PROGO elucidated the mechanistic effects of these mutations. Finally, gram-scale preparative bioconversions were performed using purified PROGO and *E. coli* cells expressing PROGO to demonstrate the practical utility of the engineered biocatalyst. Enzyme kinetics were determined using spectrophotometric assays monitoring absorbance changes.

The researchers successfully engineered a bacterial eugenol oxidase (EUGO) to create a highly active and selective biocatalyst (PROGO) for the dehydrogenation of 4-n-propylguaiacol to isoeugenol. The engineering involved a multi-step process: 1) Thermostability enhancement using FRESCO, resulting in the EUGO5X variant with a 13.5 °C increase in Tm compared to wild-type EUGO. 2) Chemoselectivity improvement by identifying the S394V mutation which increased isoeugenol production and reduced undesired byproducts. 3) Activity enhancement by addressing the formation of a covalent substrate-flavin adduct, resulting in a 100-fold increase in kcat compared to wild-type EUGO on 4-n-propylguaiacol. The final PROGO variant, comprising eight mutations (S81H, A423M, H434Y, S518P, I445D, S394V, D151E, and Q425S), demonstrated high thermostability (Tm of 81.6°C), exceptional chemoselectivity (97% isoeugenol), and significantly improved activity (kcat of 0.43 s⁻¹ for 4-n-propylguaiacol). Gram-scale bioconversions, utilizing both purified PROGO and whole cells expressing PROGO, successfully produced isoeugenol with good yields (66% and 42%, respectively). Crystal structures at different stages of engineering provided crucial insights into the structural basis of the mutations' effects. The S394V mutation reduced water accessibility in the active site, increasing chemoselectivity, while the D151E and Q425S mutations reduced the formation of the rate-limiting covalent adduct.

The successful engineering of EUGO into PROGO demonstrates the potential of a combined computational and experimental strategy for designing efficient biocatalysts for lignin valorization. The stepwise approach, guided by structural and mechanistic insights, allowed for the precise modification of enzyme properties, addressing challenges in activity, selectivity, and stability. The significant increase in kcat for 4-n-propylguaiacol conversion by PROGO showcases the effectiveness of targeting the rate-limiting covalent adduct formation. The use of both purified enzyme and whole cells expressing PROGO in gram-scale reactions validates the practical applicability of this engineered biocatalyst for industrial settings. The study highlights the power of combining computational tools with experimental validation and structural analysis to achieve targeted enzyme modifications. The methodology presented here could be applied to engineer other enzymes for the valorization of other lignin-derived compounds.

This study successfully engineered a bacterial eugenol oxidase (EUGO) into a highly efficient and selective 4-n-propylguaiacol oxidase (PROGO) for the production of isoeugenol from a lignin-derived compound. PROGO combines high thermostability, chemoselectivity, and catalytic activity, demonstrated through gram-scale bioconversions. This work showcases the potential of structure-guided enzyme engineering for converting renewable lignin feedstocks into valuable products. Future research could focus on further optimizing PROGO for industrial applications and exploring its applicability to other lignin-derived compounds.

While PROGO demonstrates significant improvement in activity and selectivity, further optimization might be possible. The lower isolated yield in whole-cell conversions compared to purified enzyme conversions highlights the need for optimizing the product isolation procedure to improve recovery. The study focused on a specific lignin-derived compound; the applicability of the engineering strategy to other substrates needs further investigation. The computational predictions, although successful in this case, are not always perfectly predictive, and experimental validation remains crucial.