Oleaginous yeasts are a group of microorganisms that possess the ability to store significant amounts of triacylglycerols (TAGs) as lipid bodies within their cells. Lipid bodies, also called lipid droplets or oil bodies, are specialized organelles found in eukaryotic cells. They serve as storage compartments for mainly TAGs (Renne, Klug et al. 2020). Oleaginous yeasts excel at accumulating lipids and can amass levels exceeding 20% of their cell mass (Ageitos, Vallejo et al. 2011) (Passoth, Brandenburg et al. 2023).

Carotenoids are naturally occurring pigments that exhibit potent antioxidant activity (Kaczor and Baranska 2016). They are known for their ability to scavenge harmful free radicals, protecting cells and tissues from oxidative damage (Fiedor and Burda 2014) (Paul, Kumari et al. 2023). Additionally, carotenoids have numerous industrial applications, including their use as natural food colorants, nutraceuticals, and additives in cosmetic and pharmaceutical products (Paul, Kumari et al. 2023) (Rapoport, Guzhova et al. 2021).

Among oleaginous yeasts, species of Rhodotorula possess a competitive edge in terms of their growth rates, as they can rapidly multiply under favorable conditions. Moreover, their metabolic flexibility allows them to efficiently utilize diverse carbon sources for lipid synthesis (Lamers, van Biezen et al. 2016) (Passoth, Brandenburg et al. 2023). Recent investigation showed that there is a considerable variability between strains of these species in terms of utilizing carbon sources and to form lipids from them (Brandenburg, Blomqvist et al. 2021) (Chmielarz, Blomqvist et al. 2021). The saprophytic species Rhodotorula toruloides is ubiquitous and can be present in habitats with broad geographical variability including decaying conifer wood in Sweden (R. toruloides CBS 14 (Rennerfelt 1937)), or Japanese soil (R. toruloides CBS 349 (Okunuki 1931)). Within the species, two distinct haplotypes exist, commonly referred to as mating types A1 and A2. CBS 14 belongs to mating type A1, while CBS 349 belongs to the A2 mating type (Sambles, Middelhaufe et al. 2017) (Cavelius, Engelhart-Straub et al. 2023). The hyphal conjugation of the A1 mating strain CBS 14 and A2 mating strain CBS 349 resulted in the formation of the type strain CBS 6016 (Banno 1967).

Major carotenoids produced by Rhodotorula species include β-carotene, γ-carotene, torulene, and torularhodin (Buzzini, Innocenti et al. 2007) (Perrier, Dubreucq et al. 1995). Among these, torulene and torularhodin are of high interest due to their potential applications in various industries. They possess higher antioxidant properties than β-carotene and can be used for feedstock, food, and cosmetic additives (Naz, Ullah et al. 2023) (Kot, Błażejak et al. 2018).

To analyse the lipid and carotenoid profiles of the yeast cells, it is crucial to employ an accurate extraction method. Among the different extraction techniques available, supercritical carbon dioxide (SC-CO₂) extraction has emerged as a notable method for effective extraction of thermolabile compounds like lipids and carotenoids (Sahena, Zaidul et al. 2009) (Saini and Keum 2018). Supercritical carbon dioxide refers to carbon dioxide gas that is maintained at a temperature and pressure beyond its critical point, where it exhibits both gas and liquid-like properties (Sihvonen, Järvenpää et al. 1999). When SC-CO₂ comes into contact with yeast cells containing lipids and carotenoids, it acts as a hydrophobic solvent to selectively extract these compounds (Shi, Mittal et al. 2007). This approach has received significant attention due to its efficiency in extracting these compounds (Tzima, Georgiopoulou et al. 2023). Also, the SC-CO₂ extraction, being non-toxic and obviating the requirement for potentially hazardous organic solvents, is well-suited for applications related to both food and feed (Rozzi and Singh 2002). Additionally, SC-CO₂ extraction can be operated at lower temperatures compared to classical extraction methods and is running in absence of oxygen, which helps preserving the chemical attributes of the extracted lipids and carotenoids (Reverchon, Donsi et al. 1993) (Sookwong and Mahatheeranont 2017). Recently, we could demonstrate that SC-CO₂ enabled extraction of carotenoids without co-extraction of lipids. Due to this, saponification could be avoided, enabling reliable quantification of torularhodin and torulene (Nagaraj, Burkina et al. 2022, Nagaraj, Blomqvist et al. 2025). We observed a higher proportion of unsaturated fatty acids in lipids extracted from R. toruloides, compared to classical extraction by Folch, indicating better protection of double bounds from oxidation (Nagaraj, Blomqvist et al. 2025).

The objective of this study was to analyse and compare the type-strain CBS 6016 of R. toruloides with its parental strains (CBS 14 and CBS 349) in order to identify strain specific differences in carotenoid and lipid formation, and to determine genetic traits correlating with the observed phenotypes. To accomplish this, we used supercritical carbon dioxide (SC-CO₂) extraction methods for extracting both lipids and carotenoids from the strains. The extracted compounds were then subjected to composition analysis using gas chromatography (GC) for lipids and ultra-high-performance liquid chromatography (UHPLC) for carotenoids. We further sequenced all three strains, performed genome assembling, and annotation to identify genes associated to fatty acid metabolism and carotenoid biosynthesis. This investigation may provide insights into differences in lipid and carotenoid production among the strains, thereby contributing to the understanding and optimization of lipid and carotenoid production in R. toruloides.

Inoculum preparation

The CBS 14, CBS 349 and CBS 6016^T strains of R. toruloides were obtained from the Westerdijk Fungal Biodiversity Institute, Utrecht, the Netherlands. Cells were stored in frozen stocks at − 80 °C. The inoculum for all three was prepared as described before by Nagaraj et al., 2022. In summary, cells from YPD-agar plates were introduced into 300 mL of YPD medium in a 3 L baffled Erlenmeyer flask. This was then incubated at 25°C for 48–72 hours at a speed of 150 rpm. Subsequently, the cells were harvested, washed with sterile saline solution, resuspended in saline and finally, inoculated into the fermenters.

Cultivation in bioreactors

The cells from the inoculum were inoculated into 1.5 L of growth medium in Minifors 2, Bench-Top bioreactors (INFORS HT, Switzerland, working volume 2 L). The growth medium used for cultivation of all three strains consisted of sterile glucose solution 70 g/l and filter-sterilized mixture of 1.7 g/l YNB (yeast nitrogen base without amino acids and ammonium sulphate, Difco™, Becton Dickinson and Company, USA), ammonium sulfate 2 g/l and yeast extract 0.75 g/l. The parameters for the cell cultivation in the bioreactors were as same as described by Nagaraj et al., 2022. After 96 hours, the cultivation was stopped. Subsequently, the cells were harvested, washed, subjected to French pressing, freeze-dried, and stored at -20°C until SC-CO₂ extraction. Samples were collected at both start point and at end point of cultivation to determine the cell dry weight.

Analytical techniques

Cell Dry Weight determination

For dry weight determination, 2 mL of the broth culture was centrifuged at 15,000 g for 2 min. The cell pellets were washed thrice with deionized water and transferred onto a pre-weighed aluminium plate, which was then dried at 105 °C for 24 h, followed by weight measurement. All samples were analysed in triplicate.

Supercritical carbon-dioxide extraction

SC-CO₂ extractor

All the extractions were made in a supercritical extractor (Jasco Supercritical Extractor, SFE 4000 series, Kovalent AB, Italy), which constitutes of a high-pressure CO₂ pump, modifier pump and a backpressure regulator for a stable and trouble-free constant flow of CO₂. The extraction vessel was placed in an air-conditioned oven to attain desired temperature. The freeze-dried samples were loaded into the extraction vessel, and the extracts were collected in the automatic extract collectors. The process was controlled using the software ChromNAV, version 2.3C.

Lipid extraction

The lipids from the freeze-dried biomass were extracted using a similar method as described by (Milanesio, Hegel et al. 2013) with some modifications. In a concise summary, the freeze-dried sample was combined with silica beads (SiLibeads Typ ZSA 2.2–2.5 mm) and loaded into the extraction vessel at a yeast cell to bead ratio of 3:2 (w/w). The extraction process was carried out at a pressure of 300 bar, a temperature of 45°C, and a CO₂ flow rate of 2 ml/min. The total duration of the extraction procedure was 180 minutes. Subsequently, the residue left after lipid extraction was used for carotenoid extraction. The extracted lipids were collected in pre-weighed brown bottles and stored at -20°C for future analysis (Nagaraj, Blomqvist et al. 2025).

Fatty acid analysis

The extracted lipid samples were methylated, and the resulting fatty acid methyl esters (FAME) were analysed using a gas chromatography (GC) system (CP-3800, CTC Analytics AG, Switzerland) fitted with a 50 m long × 22 mm i.d., 0.25 µm film thickness, BPX 70 fused-silica capillary column. The methylation was carried out using the procedure with boron trifluoride reagent described previously (Nagaraj, Burkina et al. 2022).

Carotenoid extraction

The carotenoids from the biomass after lipid extraction were extracted using a method similar to that described by Lim et al., (Lim, Lee et al. 2002) with some modifications. Along with supercritical carbon dioxide, ethanol (99.5% v/v) was introduced as a co-solvent for carotenoid extraction. The carotenoid extraction was performed at a pressure of 300 bar, a temperature of 50°C, a CO₂ flow rate of 2 ml/min, and a co-solvent flow rate of 0.2 ml/min. The extraction process lasted for a total of 180 min. The extracts were then stored at -18°C until further analysis in brown bottles to protect the carotenoids from light (Nagaraj, Blomqvist et al. 2025).

Carotenoid profile analysis using UHPLC-PDA

The analysis was conducted using a Shimadzu UHPLC-Nexera instrument (Kyoto, Japan), which consisted of an autosampler (SIL-20AC), quaternary pumps (LC-20AD), a column oven (CTO-20AC), and a PDA detector (Shimadzu, model SPD-M20A) connected in series. Control of the instruments, data acquisition, and processing were performed using LabSolutions software. Carotenoid separation was performed using an analytical RP C18 Kinetex 100 column (100 mm length, 4.6 mm internal diameter, 2.6 µm particle size; Phenomenex) and binary gradient system with the mobile phases A, consisting of acetonitrile–methanol (7:3, v/v), and B, consisting of ultrapure water with 0.1% formic acid. The gradient was set as follows: 0–3 min 60% B; 3–7 min 100% B; 7–30 min 100% B, and 30–35 min 60% B. The flow-rate was 0.3 mL/min, the column temperature was 40 °C, and the sample volume was 20 µL.

UV-visible spectra were obtained in the range of 250 to 600 nm. Specific wavelengths were selected for the detection of individual carotenoids: β-carotene and torularhodin peaks were acquired at 450 nm, ɣ-carotene at 462 nm, and torulene at 478 nm (Nagaraj, Burkina et al. 2022). The PDA detector allowed for the measurement of absorbance across the selected wavelengths, enabling the identification and quantification of the carotenoids of interest.

Genome analysis of CBS 14, CBS 349, and CBS 6016^T

DNA extraction, sequencing, genome assembly and annotation was performed as described (Martín-Hernández, Müller et al. 2022). The KEGG annotation tool was used for assigning KEGG Orthology (KO) identifiers to amino acid sequences inferred from annotated coding sequences using BlastKOALA (Kanehisa, Sato et al. 2016). The affiliation to and reconstruction of lipid and carotenoid metabolic pathway was performed using KEGG Mapper version 5 (Kanehisa, Sato et al. 2022) (Kanehisa and Sato 2020). Sequence alignments were performed using MUSCLE v5.1 using the standard muscle algorithm (PPP) (Edgar 2004) implemented in Geneious Prime v2025.1.2 (www.geneious.com). Protein domains and signatures were analysed using InterPro (Blum, Andreeva et al. 2025). The respective genome FASTA file (.fasta), annotation file (.gtf), protein sequence file (.pep), and coding sequence file (.cds) are provided as supplementary files.

Cell dry weight and lipid content

To assess growth performance, the cell dry weight and lipid content were analysed at the beginning and at the end of the cultivation. Interestingly, the hybrid strain CBS 6016^T showed a similar increase in cell biomass as CBS 14, whereas it differed significantly from the other parental strain CBS 349 (p ≤ 0.05) (Fig. 1). This was also observed in terms of lipid production at the end of the cultivation. CBS 6016^T and CBS 14, exhibited comparable lipid levels of 7.46 ± 0.5 and 6.48 ± 0.04 g/L, respectively whereas CBS349 showed significantly lower lipid production (2.17 ± 0.31 g/L).

Fatty acid profile

At the end of cultivation (96 hours), the fatty acid profile of the strains was analysed (Table 1). In all three strains of R. toruloides, oleic acid (C18:1(n-9)) was the major fatty acid followed by palmitic acid (C16:0) and linoleic acid (C18:2(n-6)), in accordance with previous observations (Nagaraj, Burkina et al. 2022) (Nagaraj, Blomqvist et al. 2025). Linoleic acid and linolenic acid (C18:3(n-3)) contributed to the prevalence of polyunsaturated fatty acids (PUFA) in the strains. Oleic acid was the most abundant monounsaturated fatty acid (MUFA), while palmitic acid was the major saturated fatty acid (SFA).

Table 1. Quantification of different fatty acids in R. toruloides strain samples at the end of cultivation. Data are presented as mean % ± standard deviation of fatty acids in the samples from three fermenters. Different superscript letters represent significant differences (p ≤ 0.05)