Oleaginous Yeast Biology Elucidated With Comparative Transcriptomics

Sarah J. Weintraub,Zekun Li,Carter L. Nakagawa,Joseph H. Collins,Eric M. Young

DOI: https://doi.org/10.1002/bit.28891

IF: 3.8

2024-12-12

Biotechnology and Bioengineering

Abstract:Non‐model extremophilic yeasts such as Debaryomyces spp. exhibit metabolic and stress‐tolerance traits valuable for biomanufacturing. This study links genotype to phenotype by comparing phenotypic and transcriptomic responses under nitrogen starvation, iron starvation, and salt stress. Using complementary computational approaches—network mapping and coexpression clustering—the study highlights mechanisms like lipid biosynthesis and riboflavin overflow metabolism in Debaryomyces. These findings provide insights into stress responses and metabolic regulation, advancing the utility of yeasts for sustainable biotechnology. Extremophilic yeasts have favorable metabolic and tolerance traits for biomanufacturing‐ like lipid biosynthesis, flavinogenesis, and halotolerance – yet the connection between these favorable phenotypes and strain genotype is not well understood. To this end, this study compares the phenotypes and gene expression patterns of biotechnologically relevant yeasts Yarrowia lipolytica, Debaryomyces hansenii, and Debaryomyces subglobosus grown under nitrogen starvation, iron starvation, and salt stress. To analyze the large data set across species and conditions, two approaches were used: a "network‐first" approach where a generalized metabolic network serves as a scaffold for mapping genes and a "cluster‐first" approach where unsupervised machine learning co‐expression analysis clusters genes. Both approaches provide insight into strain behavior. The network‐first approach corroborates that Yarrowia upregulates lipid biosynthesis during nitrogen starvation and provides new evidence that riboflavin overproduction in Debaryomyces yeasts is overflow metabolism that is routed to flavin cofactor production under salt stress. The cluster‐first approach does not rely on annotation; therefore, the coexpression analysis can identify known and novel genes involved in stress responses, mainly transcription factors and transporters. Therefore, this work links the genotype to the phenotype of biotechnologically relevant yeasts and demonstrates the utility of complementary computational approaches to gain insight from transcriptomics data across species and conditions.

biotechnology & applied microbiology