Biocatalysis, using enzymes for organic synthesis, has emerged as powerful tool for the synthesis of active pharmaceutical ingredients (APIs). The first industrial biocatalytic processes launched in the first half of the last century exploited whole-cell microorganisms where the specific enzyme at work was not known. In the meantime, novel molecular biology methods, such as efficient gene sequencing and synthesis, triggered breakthroughs in directed evolution for the rapid development of process-stable enzymes with broad substrate scope and good selectivities tailored for specific substrates. To date, enzymes are employed to enable shorter, more efficient, and more sustainable alternative routes toward (established) small molecule APIs, and are additionally used to perform standard reactions in API synthesis more efficiently. Herein, large-scale synthetic routes containing biocatalytic key steps toward >130 APIs of approved drugs and drug candidates are compared with the corresponding chemical protocols (if available) regarding the steps, reaction conditions, and scale. The review is structured according to the functional group formed in the reaction.

Cytochrome P450 CYP153A M.aq from Marinobacter aquaeolei serves as a model enzyme for the terminal (ω-) hydroxylation of medium- to long-chain fatty acids. We have engineered this enzyme using different mutagenesis approaches based on structure-sequence-alignments within the 3DM database and crystal structures of CYP153A M.aq and a homologue CYP153A P.sp . Applying these focused mutagenesis strategies and site-directed saturation mutagenesis, we created a variant that ω-hydroxylates octanoic acid. The M.aqRLT variant exhibited 151-fold improved catalytic efficiency and showed strongly improved substrate binding (25-fold reduced K m compared to the wild type). We then used molecular dynamics simulations to gain deeper insights into the dynamics of the protein. We found the tunnel modifications and the two loop regions showing greatly reduced flexibility in the engineered variant were the main features responsible for stabilizing the enzyme–substrate complex and enhancing the catalytic efficiency. Additionally, we showed that a previously known fatty acid anchor (Q129R) interacts significantly with the ligand to hold it in the reactive position, thereby boosting the activity of the variant M.aqRLT toward octanoic acid. The study demonstrates the significant effects of both substrate stabilization and the impact of enzyme flexibility on catalytic efficiency. These results could guide the future engineering of enzymes with deeply buried active sites to increase or even establish activities toward yet unknown types of substrates.

The interconversion of non-activated alkenes and alcohols, catalysed by (de)hydratases, has great potential in biotechnology for the generation of fine and bulk chemicals. LinD is a cofactor-independent enzyme that catalyses the reversible (de)hydration of the tertiary alcohol (S)-linalool to the triene b-myrcene, and also its isomerization to the primary alcohol geraniol. Structure-informed mutagenesis of LinD, followed by activity studies, confirmed essential roles for residues C171, C180 and H129 in water activation for the hydration of b-myrcene to linalool. However, no evidence of covalent thioterpene intermediates was found using either X-ray crystallography, mass spectrometry, or QM/MM nudged elastic band simulations. Labelling and NMR experiments confirmed a role for residue D39 in (de)protonation of the linalool carbon C10 in the isomerization of linalool to geraniol and also the intermediacy of b-myrcene in this isomerization reaction. X-ray, molecular dynamics and activity studies also suggested a significant role in catalysis for a mobile methionine residue M125, which exists in substantially altered orientations in different mutant structures.

No abstract