Development of a codon-optimised AUR1 selectable marker for genetic engineering of diverse yeast species

Abstract

Genetic engineering of non-conventional yeasts is frequently limited by the lack of robust dominant selection systems that function across phylogenetically diverse hosts. These organisms are increasingly important platforms for sustainable bioproduction due to their unique metabolic capabilities. Candida famata , an industrial riboflavin overproducer belonging to the CTG clade with alternative codon decoding, represents a particularly relevant model for evaluating cross-species selection strategies. Many commonly used antibiotic resistance markers exhibit strong host dependence, especially in yeasts with non-standard genetic codes, thereby restricting strain construction and metabolic engineering. Aureobasidin A (AbA) is a potent antifungal compound that inhibits inositol phosphorylceramide (IPC) synthase, a key enzyme in sphingolipid biosynthesis, making it an attractive candidate for dominant selection. In this study, we evaluated the sensitivity to AbA across phylogenetically diverse yeast species and developed a resistance marker based on a codon-optimised variant of the Saccharomyces cerevisiae AUR1 gene, designed to ensure correct translation in CTG-clade yeasts. Sensitivity assays confirmed that AbA efficiently inhibited the growth of multiple yeast species at low concentrations, supporting its use as a selective agent. The native C. albicans AUR1 gene did not confer functional resistance in C. famata despite successful transformation, highlighting limitations imposed by host-dependent gene expression. In contrast, the codon-optimized S. cerevisiae AUR1 construct restored robust resistance across the tested yeast species. Transformation yielded stable mutants with reproducible resistance phenotypes, confirmed by molecular validation and maintained through serial passaging. Importantly, the construct remained functional not only in CTG-clade yeasts but also in species using the standard genetic code. Codon usage analysis showed that replacement of CTG codons eliminated ambiguous decoding while introducing leucine codons broadly preferred among yeast species, providing a plausible explanation for the observed cross-species functionality of the AUR1 marker. Fusion with GFP confirmed correct expression and intracellular localisation without detectable impact on host physiology, while fermentation experiments demonstrated that the system did not significantly affect riboflavin production under the tested conditions. Collectively, these results establish AbA as a highly effective selective agent and demonstrate that rational recoding of AUR1* enables the development of an efficient dominant selection marker across the yeast species evaluated in this study. This platform expands the genetic toolkit available for non-conventional yeasts and provides a versatile solution for strain engineering across multiple yeast hosts.