Engineering of hydroxymandelate synthases and the aromatic amino acid pathway enables de novo biosynthesis of mandelic and 4-hydroxymandelic acid with Saccharomyces cerevisiae.

Mandelic acid (MA) and 4-hydroxymandelic acid (HMA) are valuable specialty chemicals used as precursors for flavors as well as for cosmetic and pharmaceutical purposes. Today they are mainly synthesized chemically. Their synthesis through microbial fermentation would allow for environmentally sustainable production. In this work, we engineered the yeast Saccharomyces cerevisiae for high-level production of MA and HMA. Expressing the hydroxymandelate synthase from Amycolatopsis orientalis in a yeast wild type strain resulted in the production of 119mg/L HMA from glucose. As the enzyme also accepts phenylpyruvate as a substrate aside from its native substrate 4-hydroxyphenylpyruvate, 0.7mg/L MA was also produced. Preventing binding of 4-hydroxyphenylpyruvate to the hydroxymandelate synthase by introducing a S201V replacement in its substrate binding site nearly completely prevented HMA production but increased MA production only 3.5-fold. To further increase HMA and MA production, the aromatic amino acid pathway was engineered. We increased the precursor supply by introducing modifications in the shikimic acid pathway (ARO1↑, ARO3K222L ↑, ARO4K220L ↑) and reducing flux into the Ehrlich pathway (aro10Δ), and thereby enhanced the HMA titer to 465mg/L and the MA titer to 2.9mg/L. A further increase in HMA and MA titers was achieved by replacing the hydroxymandelate synthase from A. orientalis with the corresponding enzyme from Nocardia uniformis. Subsequently, we introduced additional deletions to block the competing tryptophan branch (trp2Δ), to further decrease flux into the Ehrlich pathway (pdc5Δ) and to avoid transamination of phenylpyruvate and 4-hydroxyphenylpyruvate (aro8Δ, aro9Δ). We achieved more than 1g/L 4-hydroxymandelate when additionally preventing formation of phenylpyruvate by deleting PHA2. When deleting TYR1 to prevent formation of 4-hydroxyphenylpyruvate instead, an MA titer of 236mg/L was achieved. This is a more than 200-fold increase in MA production compared to the wild type strain expressing the hydroxymandelate synthase from A. orientalis. Finally, we showed that S. cerevisiae tolerates HMA and MA to concentrations as high as 3g/L and 7.5g/L, respectively. Our results demonstrate that S. cerevisiae is a promising host for sustainable MA and HMA production.