Site-saturation mutagenesis was used to generate all possible replacements for Trp 116 of Saccharomyces pastorianus (formerly Saccharomyces carlsbergensis) old yellow enzyme (OYE). Our original hypothesis-that smaller amino acids at position 116 would allow better acceptance of bulky 3-alkyl-substituted 2-cyclohexenones-proved incorrect. Instead, Phe and Ile replacements favored the binding of some substrates in an opposite orientation, which yielded reversed stereochemical outcomes compared to that of the wild-type OYE. For example, W1161 OYE reduced (R)- and (S)-carvone to enantiomeric products, rather than the diastereomers produced by the wild-type OYE. Deuterium labeling revealed that (S)-carvone reduction by the W1161 OYE occurred by the same pathway as that by the wild type (net trans-addition of H-2), proving that different substrate binding orientations were responsible for the divergent products. Trp 116 mutants also afforded different stereochemical outcomes for reductions of (R)perillaldehyde and neral. Preliminary studies of an OYE family member whose native sequence contains Ile at position 116 (Pichia stipitis OYE 2.6) revealed that this enzyme's stereoselectivity matched that of the wild-type S. pastorianus OYE, showing that the identity of the residue at position 116 does not solely determine the substrate binding orientation. Computational docking studies using an induced fit methodology successfully reproduced the majority of the experimental outcomes. These computational tools will allow preliminary in silico screening of additional residues to identify those most likely to control the substrate binding orientation and provide some guidance to future experimental studies.