Saccharomyces cerevisiae strains performing similarly during fermentation of lignocellulosic hydrolysates show pronounced differences in transcriptional stress responses

Abstract

Improving our understanding of the transcriptional changes of Saccharomyces cerevisiae during fermentation of lignocellulosic hydrolysates is crucial for the creation of more efficient strains to be used in biorefineries. We performed RNA sequencing of a CEN.PK laboratory strain, two industrial strains (KE6-12 and Ethanol Red), and two wild-type isolates of the LBCM collection when cultivated anaerobically in wheat straw hydrolysate. Many of the differently expressed genes identified among the strains have previously been reported to be important for tolerance to lignocellulosic hydrolysates or inhibitors therein. Our study demonstrates that stress responses typically identified during aerobic conditions such as glutathione metabolism, osmotolerance, and detoxification processes also are important for anaerobic processes. Overall, the transcriptomic responses were largely strain dependent, and we focused our study on similarities and differences in the transcriptomes of the LBCM strains. The expression of sugar transporter-encoding genes was higher in LBCM31 compared with LBCM109 that showed high expression of genes involved in iron metabolism and genes promoting the accumulation of sphingolipids, phospholipids, and ergosterol. These results highlight different evolutionary adaptations enabling S. cerevisiae to strive in lignocellulosic hydrolysates and suggest novel gene targets for improving fermentation performance and robustness.

Importance: The need for sustainable alternatives to oil-based production of biochemicals and biofuels is undisputable. Saccharomyces cerevisiae is the most commonly used industrial fermentation workhorse. The fermentation of lignocellulosic hydrolysates, second-generation biomass unsuited for food and feed, is still hampered by lowered productivities as the raw material is inhibitory for the cells. In order to map the genetic responses of different S. cerevisiae strains, we performed RNA sequencing of a CEN.PK laboratory strain, two industrial strains (KE6-12 and Ethanol Red), and two wild-type isolates of the LBCM collection when cultivated anaerobically in wheat straw hydrolysate. While the response to inhibitors of S. cerevisiae has been studied earlier, this has in previous studies been done in aerobic conditions. The transcriptomic analysis highlights different evolutionary adaptations among the different S. cerevisiae strains and suggests novel gene targets for improving fermentation performance and robustness.

Keywords: RNA sequencing; industrial yeast strains; inhibitor stress; tolerance; wild-type isolates.

Fig 1

Anaerobic cultivation of CEN.PK113-7D (red squares), KE6-12 (yellow circles), Ethanol Red (green triangles), LBCM31 (blue diamonds), and LBCM109 (purple crosses) in minimal medium containing 70% WSH. Sampling time for each culture is indicated by the vertical dashed line in the corresponding color. Data obtained from four biological replicates; shadows show the standard deviation.

Fig 2

Unsupervised multi-dimensional scaling plot of all RNA sequencing samples of CEN.PK113-7D (red squares), KE6-12 (yellow circles), Ethanol Red (green triangles), LBCM31 (blue diamonds), and LBCM109 (purple crosses). X and Y axes represent the first (dim 1) and second (dim 2) leading fold change that best separates the samples and explains the largest proportion of variation in the data.

Fig 3

Counts of the significant DEGs between the strains analyzed. The number of genes that were expressed at a significantly higher (red bars) or lower (green bars) level for the strain reported at the bottom of the bars compared with the one specified at the top; CEN.PK113-7D, KE6-12, Ethanol Red, and LBCM31. Significance was defined as adjusted P value < 0.01 and fold change ≥ 2. Data presented are based on the average of four biological replicates.

Fig 4

Expression of genes related to glutathione metabolism and NADPH regeneration. (a) Schematic map depicting the metabolic pathway of glutathione according to the KEGG pathway representation. Elements in pink and blue represent genes that are expressed at a significantly (adjusted P value <  0.01) higher or lower level in both LBCM strains compared with CEN.PK113-7D. (b) Differential expression of genes related to glutathione metabolism in LBCM31 (blue bars) and LBCM109 (purple bars) compared with CEN.PK113-7D. The relative expression level of each gene is visualized as log2 of the fold change (log2 FC). The letters “ns” above the last bar represent the statistically non-significant (adjusted P value >  0.01) differential gene expression for that comparison.

Fig 5

GO term analysis of genes differently expressed in LBCM109 when compared with LBCM31. Percentages of expressed genes at a significantly (adjusted P value < 0.01) higher or lower level are marked in dark-red or dark-blue, respectively, whereas genes differently expressed, although not at a significant level (adjusted P value > 0.01), are marked in light-red or light-blue, respectively. The name of each GO term is inside the left or the right side of its relative bar, depending on whether the majority of the genes for that GO term are expressed at a higher (left side) or lower (right side) level. Each GO term name is followed by the total number of genes of that GO term. Data obtained from four biological replicates.

Fig 6

Log₂ FC of the (a) 10 genes expressed at highest or lowest level; (b) most significantly differentially expressed genes of the “lipid metabolism” GO term; (c) most significantly differentially expressed genes involved in ergosterol biosynthesis; (d) most significantly differentially expressed genes involved in iron metabolism or sugar transport in LBCM109 when compared with LBCM31. Significance was defined as adjusted P value < 0.01. Data are obtained from four biological replicates. All data on the DEGs are found in Tables S3-S4.