The aryl-alcohol oxidase (AAO) is a fungal flavoenzyme that supplies H2O2 to the ligninolytic consortium during natural wood decay. Being active on a wide array of aromatic alcohols, this GMC oxidase presents a highly enantioselective mechanism of great interest in organic synthesis processes. The most powerful strategy for the AAO to meet industrial standards is the engineering of its properties by directed evolution. In the present Doctoral Thesis, an evolutionary platform for the AAO from Pleurotus eryngii was developed in order to: (i) obtain functional expression in yeasts, (ii) design a secondary benzyl-alcohol oxidase, and (iii) explore the enzymatic conversion of furfural derivatives. To achieve functional expression in Saccharomyces cerevisiae, the AAO gene was fused to different signal peptides including chimeric versions of the mating-α factor and the killer K1 toxin preprosequences. The platform for in vitro evolution was completed with a dual high-throughput screening assay to detect H2O2 that included a method based on the Fenton reaction. To enhance secretion, several libraries were created combining classical evolution (i.e. mutagenic PCR and DNA shuffling) with structure-guided evolution by MORPHING. The final secretion variant FX9, carried four mutations in the signal peptide and two substitutions in the mature protein including the consensus/ancestral H91N. The FX9 improved secretion up to 4.5 mg/L and presented high stability and kinetic values similar to the native enzyme. FX9 was cloned and expressed in Pichia pastoris maintaining expression levels and main biochemical properties. When the production was scaled-up in 5L fermenter, AAO production was increased to 25.5 mg/L. FX9 was further evolved to selectively oxidize secondary benzyl alcohols. The residual activity on chiral molecules was unlocked with the modulation for the catalytic pocket by combinatorial saturation mutagenesis. After four generations, that included a site-directed recombination step to polish mutations, LanDo variant harbored five new substitutions increasing the catalytic efficiency with 1-(p-methoxyphenyl)-ethanol in 3 orders of magnitude with a 99% ee. Exploring the transformation of 5-hydroxymethylfurfural (HMF) into furan-2,5-dicarboxylic acid (FDCA), FX9 acquired mutation F501W that improved catalytic efficiency on HMF 3-fold and showed for the first time the performance of three consecutive oxidations for the AAO.