RESUMO
Many enzymes have latent activities that can be used in the conversion of non-natural reactants for novel organic conversions. A classic example is the conversion of benzaldehyde to a phenylacetyl carbinol, a precursor for ephedrine manufacture. It is often tacitly assumed that purified enzymes are more promising catalysts than whole cells, despite the lower cost and easier maintenance of the latter. Competing substrates inside the cell have been known to elicit currently hard-to-predict selectivities that are not easily measured inside the living cell. We employ NMR spectroscopic assays to rationally combine isomers for selective reactions in commercial S. cerevisiae. This approach uses internal competition between alternative pathways of aldehyde clearance in yeast, leading to altered selectivities compared to catalysis with the purified enzyme. In this manner, 4-fluorobenzyl alcohol and 2-fluorophenylacetyl carbinol can be formed with selectivities in the order of 90%. Modification of the cellular redox state can be used to tune product composition further. Hyperpolarized NMR shows that the cellular reaction and pathway usage are affected by the xenochemical. Overall, we find that the rational construction of ternary or more complex substrate mixtures can be used for in-cell NMR spectroscopy to optimize the upgrading of similar xenochemicals to dissimilar products with cheap whole-cell catalysts.