The development of new wine yeast strains with improved characteristics is critical in the highly competitive wine market, which faces the demand of ever-changing consumer preferences. Although new strains can be constructed using recombinant DNA technologies, consumer concerns about genetically modified (GM) organisms strongly limit their use in food and beverage production. We have applied a non-GM approach, adaptive evolution with sulfite at alkaline pH as a selective agent, to create a stable yeast strain with enhanced glycerol production; a desirable characteristic for wine palate. A mutant isolated using this approach produced 41% more glycerol than the parental strain it was derived from, and had enhanced sulfite tolerance. Backcrossing to produce heterozygous diploids revealed that the high-glycerol phenotype is recessive, while tolerance to sulfite was partially dominant, and these traits, at least in part, segregated from each other. This work demonstrates the potential of adaptive evolution for development of novel non-GM yeast strains, and highlights the complexity of adaptive responses to sulfite selection.