RESUMEN
Glucolipids (GLs) are glycolipid biosurfactants with promising properties. These GLs are composed of glucose attached to a hydroxy fatty acid through a ω and/or ω-1 glycosidic linkage. Up until today these interesting molecules could only be produced using an engineered Starmerella bombicola strain (∆ugtB1::URA3 G9) producing GLs instead of sophorolipids, albeit with a very low average productivity (0.01 g·L-1 ·h-1 ). In this study, we investigated the reason(s) for this via reverse-transcription quantitative polymerase chain reaction and Liquid chromatography-multireaction monitoring-mass spectrometry. We found that all glycolipid biosynthetic genes and enzymes were downregulated in the ∆ugtB1 G9 strain in comparison to the wild type. The underlying reason for this downregulation was further investigated by performing quantitative metabolome comparison of the ∆ugtB1 G9 strain with the wild type and two other engineered strains also tinkered in their glycolipid biosynthetic gene cluster. This analysis revealed a clear distortion of the entire metabolism of the ∆ugtB1 G9 strain compared to all the other strains. Because the parental strain of the former was a spontaneous ∆ura3 mutant potentially containing other "hidden" mutations, a new GL production strain was generated based on a rationally engineered ∆ura3 mutant (PT36). Indeed, a 50-fold GL productivity increase (0.51 g·L-1 ·h-1 ) was obtained with the new ∆ugtB1::URA3 PT36 strain compared with the G9-based strain (0.01 g·L-1 ·h-1 ) in a 10 L bioreactor experiment, yielding 118 g/L GLs instead of 8.39 g/L. Purification was investigated and basic properties of the purified GLs were determined. This study forms the base for further development and optimization of S. bombicola as a production platform strain for (new) biochemicals.