BACKGROUND: Yarrowia lipolytica, one of the most widely studied "nonconventional" oleaginous yeast species, is unable to grow on cellobiose. Engineering cellobiose-degrading ability into this yeast is a vital step towards the development of cellulolytic biocatalysts suitable for consolidated bioprocessing.RESULTS:In the present work, we identified six genes encoding putative β-glucosidases in the Y. lipolytica genome. To study these, homologous expression was attempted in Y. lipolytica JMY1212 Zeta. Two strains overexpressing BGL1 (YALI0F16027g) and BGL2 (YALI0B14289g) produced β-glucosidase activity and were able to degrade cellobiose, while the other four did not display any detectable activity. The two active β-glucosidases, one of which was mainly cell-associated while the other was present in the extracellular medium, were purified and characterized. The two Bgls were most active at 40-45°C and pH 4.0-4.5, and exhibited hydrolytic activity on various β-glycoside substrates. Specifically, Bgl1 displayed 12.5-fold higher catalytic efficiency on cellobiose than Bgl2. Significantly, in experiments where cellobiose or cellulose (performed in the presence of a β-glucosidase-deficient commercial cellulase cocktail produced by Trichoderma reseei) was used as carbon source for aerobic cultivation, Y. lipolytica ∆pox co-expressing BGL1 and BGL2 grew better than the Y. lipolytica strains expressing single BGLs. The specific growth rate and biomass yield of Y. lipolytica JMY1212 co-expressing BGL1 and BGL2 were 0.15 h(-1) and 0.50 g-DCW/g-cellobiose, respectively, similar to that of the control grown on glucose.CONCLUSIONS:We conclude that the bi-functional Y. lipolytica developed in the current study represents a vital step towards the creation of a cellulolytic yeast strain that can be used for lipid production from lignocellulosic biomass. When used in combination with commercial cellulolytic cocktails, this strain will no doubt reduce enzyme requirements and thus costs.