In the development of S. cerevisiae for second generation ethanol production there is a continued need for high performance D-xylose metabolizing strains. Irrespective of the engineering strategy used, D-xylose metabolizing strains depend on the expression of Xks1 which converts D-xylulose in to D-xylulose-5-phosphate at the expense of one ATP. Since this is the committing step in D-xylose metabolism, regulation of XKS1 would potentially be required. However, D-xylose mediated regulation in S. cerevisiae does not appear to exist, likely because this sugar is not recognized as carbon source. Remarkably, also in the naturally D-xylose metabolizing yeast Pichia stipitis, the expression of XKS1 (or XYL3) is not regulated by the D-xylose concentration (Han et al. 2010). In contrast, the conversion of D-glucose to D-glucose-6-phosphate by hexokinase is highly regulated in S. cerevisiae through gene expression (47), protein degradation (48–50) and negative feedback loops (38, 51). A major regulatory role in D-glucose metabolism is fulfilled by Hxk2, which not only catalyzes the phosphorylation of D-glucose into D-glucose-6-phosphate, but that is also required for the glucose-induced repression of several genes, including HXK1 and GLK1, and for glucose-induced expression HXK2 itself (41, 52, 53). The rate of D-glucose phosphorylation is also determined by a negative feedback loop thereby limiting the amount of ATP being consumed at high D-glucose availability. Accumulation of D-glucose-6-phosphate results in increased D-trehalose-6-phosphate levels, produced by the trehalose pathway, which decreases through direct inhibition the phosphorylation of D-glucose by Hxk2. Therefore the deletion of Tps1, which converts D-glucose-6-phosphate into D-trehalose-6-phosphate, is lethal for strains grown on D-glucose which has been attributed to ATP depletion (42, 43). Hence, the overexpression of Xks1 combined with the absence of a negative feedback loop could, at high D-xylulose concentrations, potentially lead to rapid ATP consumption and cause substrate accelerated death. Such a phenomenon was observed with D-glucose conversion by Hxk2 which, when the negative feedback loop was deleted (TPS1), showed substrate accelerated death since all D-glucose is instantaneously converted in to D-glucose-6-phosphate thereby draining all ATP (42, 54, 55).
Increased Xks1 expression causes an increase in D-xylulose consumption in a wild-type S. cerevisiae (56). In contrast, Jin et. al. show that in their metabolically engineered S. cerevisiae strain, only moderate transcript levels of XKS1 are required for improved D-xylose consumption (57) which is in agreement with Latimer et. al. (58). Further, in a wild-type S. cerevisiae strain, overexpression of XKS1 is lethal when cells are grown solely on D-xylulose (59). These results point at a phenomenon of substrate accelerated death with D-xylose as substrate, but the hypothesis of ATP depletion was not further experimentally tested nor where conditions explored where the phenomenon does not occur. Here, we show that decreased expression of XKS1, using the galactose tunable expression system (Fig. 3), improves growth at high D-xylose concentrations (Fig. 4B) in a xylose consuming strain. In the wild-type IMX730 strain, already at 2% D-xylose, growth rate reduction could be observed (Fig. 1, 4A) which is accompanied with decreased ATP levels (Fig. 2). At higher D-xylose concentrations, these effects are further exacerbated, resulting in near to complete growth arrest at 8% D-xylose (Fig. 2). ATP levels in S. cerevisiae were previously studied under various conditions (or inhibitors) and it was shown that even during starvation (either carbon or nitrogen) the ATP levels drops only up to ~ 50% (60–64). Moreover, Takaine et. al. recently showed that the stable maintenance of ATP is essential for proteostasis and that ATP levels remain remarkably stable throughout different growth phases (65). We show that ATP levels decrease to similar levels as observed previously in starving cells, and therefore conclude that the fast conversion of D-xylulose to D-xylulose-5-phosphate by Xks1 drains the intracellular ATP levels in the cells. The current data underscores previous observations that a moderate expression of Xks1 improves the D-xylose consumption (57, 58), however in those studies a direct link to ATP depletion was not demonstrated. Wahlbom et.al. showed that XKS1 was upregulated in an evolutionary engineering experiment with a strain expressing the XK/XRD pathway and grown on 2% D-xylose (66). The abovementioned data shows that conditions, strain and pathway determine the outcome of the optimal expression level of Xks1. Thus with a defined feedback mechanism of the regulation of the metabolic flux through Xks1, external interference is necessary to realize the optimal balance between the metabolic flux and the availability of ATP. Further, engineering of the D-xylose pathway appears not nearly as efficient as glycolysis and still requires fine-tuning either in terms of gene expression, protein degradation or the engineering of a D-xylose sensing system in S. cerevisiae (67, 68). Ideally, this would mean that a demand-dependent expression of XKS1 may be employed to maintain a high flux of D-xylose metabolism during the fermentation until all D-xylose is utilized. This would require a flexible XKS1 expression system based on a genetic circuit that is composed of a promotor that senses D-xylose at low concentrations (comparable to the promoters of Hxt6/7) and a D-xylose sensing system based on e.g. Rgt2/Snf3. Both aspects, D-xylose promotors (69, 70) and D-xylose sensing (71), still requires more research and implementation in order to yield an economical feasible second generation biofuel process.