Characterization of red pigmented yeasts and genes associated with astaxanthin synthesis in Rhodotorula mucilaginosa HL26-1 and Rhodotorula paludigena LL69-1

Isolation and phenotypic characterization

We collected eleven flower samples from the residential areas in Lampang Province, Thailand, in April 2022. Flower identification was conducted on-site by interviewing the owner of the plant and was then identified by one of the co-authors (B.S.). The samples were carefully transported to the laboratory in sterile plastic bags with ice packs to maintain integrity during transit. Specimens have been deposited at the Department of Pharmacognosy and Pharmaceutical Botany, Faculty of Pharmaceutical Sciences, Chulalongkorn University, Thailand. Detailed information regarding the samples, specimen number, and their respective collection locations is provided in Table S1. Each sample was inoculated into 15 mL of yeast malt (YM) medium containing 10 g/L glucose, 5 g/L peptone, 3 g/L yeast extract, and 3 g/L malt extract at pH 5.5. The inoculated media were incubated at 25°C for 3 days. Subsequently, the resulting cultures were streaked onto YM agar medium containing 20 g/L agar, 10 g/L glucose, 5 g/L peptone, 3 g/L yeast extract, and 3 g/L malt extract (pH 5.5) and incubated at 30°C for 5 days. Following incubation, pigmented yeast colonies displaying the desired pink to red hue were meticulously selected and purified for clonality on YM agar media using the streak plate method. Each clonal culture was carefully preserved on a YM agar slant at 4°C to facilitate further analysis and experimentation. Carbon assimilation of strain was performed using the API^® ID 32 C kit (BioMérieux, France) according to the manufacturer’s instructions. Reactions were visually examined at 72 h, and then the results were interpreted to be positive or negative based on the presence or absence of turbidity in the carbohydrate wells.

Molecular identification of yeast strains

Genomic DNA extraction

Genomic DNA was extracted from the pure yeast cultures using the glass bead extraction method²² with some modification. Briefly, yeast cells were washed twice with sterilized distilled water and then lysed by vortexing with 0.3 g of ∅ 0.45-0.52 mm acid washed glass beads in 200 μL of extraction buffer (comprising 2% (v/v) Triton X-100, 1% (w/v) SDS, 100 mM NaCl, 10 mM Tris-HCl (pH 8.0), and 1 mM EDTA (pH 8.0)) for 5 minutes. The resulting cell-free supernatant was transferred to a new tube and gently mixed with a 2X volume of phenol (Sigma–Aldrich^®, USA, Cat. No. 242322), chloroform (RCI Labscan^®, Ireland, Cat. No. AR 1027E), and isoamyl alcohol (Sigma–Aldrich^®, USA, Cat. No. 8.22255) at a ratio of 25:24:1. After centrifugation, the supernatant was transferred to a new tube containing 1 mL of absolute ethanol (RCI Labscan^®, Ireland, Cat. No. AR1069) and stored at -20°C for 20 minutes. The DNA pellet obtained after centrifugation was washed with 500 μL of 70% (v/v) ethanol, dried at 37°C, dissolved in 50 μL of TE buffer, and stored at -20°C until further use.

Sequencing of the 26S rRNA gene (D1/D2 domain)

The 26S rRNA gene within the D1/D2 domain of the large subunit (LSU D1/D2 domain) was amplified by PCR using the primers NL1 (5′-GCATATCAATAAGCGGAGGAAAAG-3′) and NL4 (5′-GGTCCGTGTTTCAAGACGG-3′).²³ PCR was conducted in a 20 μL volume comprising 2 μL of DNA template, 0.4 μL of each primer (10 pmol/μL), 10 μL of 2X Go Tag Green, and 7.2 μL of distilled water. The amplification process involved an initial denaturation step at 94°C for 3 minutes, followed by 36 cycles of denaturation at 94°C, annealing at 52°C, and extension at 72°C, each for 30 seconds, and a final extension step at 72°C for 5 minutes. Subsequently, the PCR products were purified using a gel/PCR DNA fragment extraction kit (GenepHlow™, Geneaid Biotech Ltd., Taiwan). Purified PCR products were sequenced bidirectionally using BT sequencing technology (Celemics Inc., Republic of Korea).

Analysis of phylogenetic placement

The LSU D1/D2 sequences were aligned with those of related species using MUSCLE,²⁴ and any gaps were removed. MEGA11 software²⁵ was used to construct a neighbor-joining (NJ) tree using Kimura’s two-parameter model.²⁶ The reliability of the branches was assessed using a bootstrap test with 1000 replicates.²⁷

Yeast cultivation for astaxanthin production

A single loopful of yeast cultured on YM agar was transferred to YM broth (50 mL) in a 250 mL flask and then incubated at 30°C with agitation at 200 rpm for 24 hours. Subsequently, a 5 mL aliquot of this culture was inoculated into fresh YM broth (45 mL) in a 250 mL flask and incubated under the same conditions for 72 hours. Following incubation, the cells were harvested by centrifugation at 4°C and 6,500 × g, for 10 minutes, washed twice with distilled water, and then subjected to lyophilization for drying. The resultant dry cell weight (DCW) was determined to quantify the cell biomass. Lyophilized cells were used for further analysis of astaxanthin.

Astaxanthin analysis

Qualitative analysis of astaxanthin production by thin layer chromatography

The extraction and qualitative analysis of astaxanthin were performed using a slightly modified method described by Ushakumari and Ramanujan.²⁸ Lyophilized cells (0.01 g DCW) were suspended in 1 mL of acetone (RCI Labscan^®, Ireland, Cat. No. AR1003) and homogenized using a pestle at room temperature for 3 min. The supernatant was collected by centrifugation at 12,300 × g, for 10 minutes. Subsequently, a 20 μL aliquot of the extracted sample was applied to a TLC silica gel G plate using a capillary along with standard astaxanthin (DYCN^®, Dayang Chem Co. Ltd., China). A mixture of acetone (RCI Labscan^®, Ireland, Cat. No. AR1003) and hexane (RCI Labscan^®, Ireland, Cat. No. AR1085) was used as the mobile phase for spot development at a ratio of 1:3 (v/v). The colored bands were directly observed after development and compared with the standard astaxanthin band under visible light. In addition, the retention factor (R_f) value was calculated to quantify the movement of the materials along the plate.

Quantification of astaxanthin production

The extraction and measurement of astaxanthin were conducted using previously described methods with slight modifications.²⁹ To extract the intracellular carotenoid content, 50 mg of lyophilized cells were suspended in 5 mL of DMSO (Sigma–Aldrich^®, USA, Cat. No. 34869), ultrasonicated at 37 kHz and 50°C for 30 min (Elmasonic, E60H model, Germany). The resulting cell extracts were centrifuged at 12,300 × g, for 5 minutes. The extraction process was repeated until the supernatant became colorless. The astaxanthin concentration was determined using a Cary 60 UV–Vis spectrophotometer (Agilent) at 530 nm against a pure DMSO blank.^29,30 A standard curve of absorbance versus astaxanthin concentration was generated using the following concentrations: 0, 0.25, 0.5, 1, 2, 4, 6, and 8 μg/mL in DMSO. Astaxanthin concentrations were calculated using the standard calibration curve of AXs. The results are presented as the mean of triplicate measurements.

Confirmative analysis of astaxanthin synthesis

The astaxanthin was extracted and analyzed based on the described method.¹⁵ In brief, 0.05 g of dried yeast cell and 5 mL of DMSO (Sigma–Aldrich^®, USA, Cat. No. 34869) were mixed and sonicated at 55°C, 5 min with 37 kHz of sonication (Elmasonic E60H, Germany). After centrifugation at 6,500 × g for 10 min, the resultant supernatant was filtered (0.22 μm) and then was transferred to chromatography column. Astaxanthin were analyzed by high performance liquid chromatography (HPLC) using UHPLC Nexera X2 system (Shimadzu, Japan) consisting of LC-30AD binary pump with a CTO-20AC column oven, an SPD-M30A detector and an SIL-30AC autosampler. Astaxanthin were separated using C18 column (GL Science InterSustain, 4.6 mm × 150 mm, 5 μm) using mobile phases consisting of mixture of methanol (RCI Labscan^®, Ireland, Cat. No. LC1115)/acetonitrile (RCI Labscan^®, Ireland, Cat. No. LC1005)/ethyl acetate (RCI Labscan^®, Ireland, Cat. No. LC1070) /formic acid (Sigma–Aldrich^®, USA, Cat. No. 5.43804) (75.9:12:12:0.1, v/v) and methanol. Analytical-grade astaxanthin was purchased from Dayang Chem Co. Ltd. (China) and utilized as a reference standard.

Whole-genome sequencing and analysis

The whole genome was sequenced using the paired-end (PE) 150 method on the Illumina HiSeq Xten/Novaseq/MGI2000 platform at Vishuo Biomedical Pte. Ltd., Beijing, China. The single-end reads were processed to eliminate adapters and low-quality bases using Fastp (v0.23.0). The resulting data were then assembled into contigs using Velvet de Novo assembler version 1.2.10.^31,32 Subsequently, the contigs were assembled into scaffolds using SSPACE (version 3.0),³³ and the gaps were filled using GapFiller (versions 1–10).³⁴ Gene prediction was performed using Augustus version 3.3.³⁵ The coding genes were annotated using the National Center for Biotechnology Information (NCBI) NR database via BLAST. Subsequently, gene function prediction was carried out using the Kyoto Encyclopedia of Genes and Genomes (KEGG)³⁶ and KofamKOALA tools, employing default settings and considering all hits (https://www.genome.jp/kegg/).³⁷ Proksee (https://proksee.ca/) was used to generate the circular genome map, and OrthoVenn (https://orthovenn3.bioinfotoolkits.net/ start/db) was used to create the Venn diagram. The Kostas Lab web-based tool (http://enve-omics.ce.gatech.edu/) was used to analyze the average nucleotide identity (ANI) values.³⁸

Isolation and identification of yeast strains

Twenty-three yeasts were isolated from the 11 flower samples collected (Table S1). Among them, 12 isolates exhibited pigmented colonies with the desired coloration ranging from pink to orange. All isolates were subjected to identification at the molecular operational taxonomic unit (MOTU) level by sequencing of the 26S rRNA gene (LSU D1/D2 domain), followed by species assignment via comparison with entries in the NCBI GenBank database using the BLASTn program. Consequently, all 12 pigment yeast strains were identified as basidiomycetous yeasts (Table S1), with two species belonging to the Rhodotorula genus: Rhodotorula paludigena (9 strains) and Rhodotorula mucilaginosa (3 strains). The remaining 11 strains were classified into four species of Basidiomycota—Cryptococcus heveanensis (1 strain), Pseudozyma aphidis (2 strains), Pseudozyma hubeiensis (1 strain), and Pseudozyma siamensis (1 strain)—and five species of Ascomycota—Candida parapsilosis (1 strain), Metschnikowia koreensis (1 strain), Saccharomyces cerevisiae (2 strains), Wickerhamiella infanticola (1 strain), and Debaryomyces nepalensis (1 strain).

Qualitative analysis of astaxanthin production by thin layer chromatography

A total of twelve pigmented strains were selected in order to conduct a qualitative examination of their astaxanthin synthesis by using the thin-layer chromatography (TLC) technique. Astaxanthin, which is red, requires no additional processing for spot visualization. Compared with that of the astaxanthin standard, an astaxanthin band was observed in all extracts, with a R_f value of 0.28. The experimental results demonstrated the presence of an astaxanthin band in all tested yeast strains ( Figure 1). All yeast strains were selected for further quantitative analysis of astaxanthin content.

Figure 1. Thin layer chromatography (TLC) of an acetone extract of pigment isolated yeasts and astaxanthin standard on Silica gel G plate using a mixture of acetone and hexane (ratio of 1:3 (v/v) as the mobile phase.

The presence of astaxanthin band of R. paludigena strain TL35-3, HL55-2, TL35-4, and TL35-6 (a), strain LL69-1, IL45-1, IL45-2-1, HL55-1, HL55 (b), and R. mucilaginosa strain TL35-2, AL44-2, and HL26-1 (c).

Quantification of astaxanthin production

Twelve yeast strains exhibiting positive results in TLC analysis for astaxanthin were subjected to quantitative analysis using spectrophotometry. Intracellular astaxanthin was extracted from powdery cells using the conventional DMSO extraction method²⁹ coupled with sound energy via an ultrasonic bath to enhance cell disruption efficiency.³⁹ In the R. mucilaginosa strains, the astaxanthin content ranged from 23.99 ± 0.22 to 104.98 ± 0.13 μg/g DCW, with yields ranging from 0.1991 ± 0.0021 to 0.9280 ± 0.0012 mg/L ( Figure 2). Moreover, several R. paludigena strains exhibited astaxanthin contents and yields ranging from 47.99 ± 0.11 to 254.78 ± 0.27 μg/g DCW and from 0.3589 ± 0.0009 to 1.8632 ± 0.0023 mg/L, respectively ( Figure 2). Strain LL69-1 demonstrated the highest astaxanthin production of R. paludigena, with an astaxanthin content of 254.78 ± 0.27 μg/g DCW and a yield of 1.8632 ± 0.0023 mg/L. Meanwhile, R. mucilaginosa strain HL26-1, the highest producer of its species, which yielded 104.98 ± 0.13 μg/g DCW and 0.9280 ± 0.0012 mg/L.

Figure 2. Quantitative analysis of astaxanthin production by 3 R. mucilaginosa strains (AL44-2, TL35-2, and HL26-1) and 9 R. paludigena strains (TL35-5, TL35-4, TL35-6, HL55, HL55-1, HL55-2, IL45-2-1, IL45-4, and LL69-1).

The results are reported as the average value ± standard deviation (SD) calculated from three repeated measurements.

Confirmative analysis of astaxanthin synthesis in selected astaxanthin-producing strains

The ability of two astaxanthin-producing yeasts, R. mucilaginosa HL26-1 and R. paludigena LL69-1, to synthesize astaxanthin was confirmed by high-performance liquid chromatography (HPLC). The HPLC chromatogram displayed astaxanthin peaks from strains HL26-1 and LL69-1, along with a 10 ppm astaxanthin standard. The retention time of the standard peak was used as a reference, confirming the presence of astaxanthin in the strain HL26-1 and LL69-1 with accuracy. The highest points within the time intervals during which HL26-1 and LL69-1 were retained were observed at 2.061 and 2.211 minutes, respectively. These findings provided conclusive evidence that both strains, HL26-1 and LL69-1, effectively synthesized astaxanthin (Supplementary Figure S1).

General features of the R. paludigena LL69-1 and R. mucilaginosa HL26-1 genomes

The genome assembly statistics for the two astaxanthin-producing yeasts, R. mucilaginosa HL26-1 and R. paludigena LL69-1, are presented in Table 1. The assembly of R. mucilaginosa HL26-1 resulted in 154 scaffolds, with the giant scaffold spanning 1,230,191 base pairs (bp). The N50 length of the scaffolds was 476,521 bp, with a GC content of 63.82% and a genome size of 20.99 Mbp. Conversely, R. paludigena LL69-1 had a genome size of 18.78 Mbp, with a GC content of 60.12%. Its assembly comprised 107 scaffolds with an N50 value of 254,092 bp and an L50 value of 24 scaffolds. The largest scaffold in LL69-1 was 853,611 bp, whereas the shortest was 329 bp. Circular genomics of the HL26-1 and LL69-1 genomes, illustrating the open reading frame (ORF) positions, GC content, and GC skew of each strain, are depicted in Figure 3. The whole-genome sequences of R. mucilaginosa HL26-1 and R. paludigena LL69-1 were deposited in the NCBI/GenBank database (http://www.ncbi.nlm.nih.gov), associated with the BioProject identities PRJNA1025132 and PRJNA1025134, BioSample numbers SAMN37714267 and SAMN37714269, and GenBank accession numbers JAZBNE000000000 and JAWJBI000000000, respectively.

Figure 3. Circular genomic map of R. mucilaginosa HL26-1 and R. paludigena LL69-1 displaying the following information: GC content in orange, GC skew (+) in blue, GC skew (-) in purple, and open reading frames (ORFs) in green.

Phylogenetic and average nucleotide identity (ANI) analysis of strains

Figure S2 depicts a phylogenetic tree using the D1/D2 domains of the large subunit sequences obtained in this study and those available in the NCBI nucleotide database. This analysis confirmed the phylogenetic relationship between these pigmented yeasts and other related species. R. mucilaginosa HL26-1 and R. paludigena LL69-1 were clustered within the same clade as their closest-type strains, R. mucilaginosa CBS316^T and R. paludigena CBS6566^T, respectively, with 100% bootstrap support.

Moreover, we conducted an average nucleotide identity (ANI) analysis, a crucial step in assessing the phylogenomic relationships between the two astaxanthin-producing yeasts and other Rhodotorula species. The pairwise ANI values of the whole genomes of R. mucilaginosa HL26-1, R. paludigena LL69-1, and the other Rhodotorula species varied (Supplementary Figure S3).

Functional genome annotation and comparative analysis of putative gene families associated with astaxanthin biosynthesis

According to annotation results from the National Center for Biotechnology Information (NCBI) database, the draft genome sequence of R. mucilaginosa HL26-1 comprises 5,711 protein-encoding genes, while R. paludigena LL69-1 was predicted to contain approximately 6,782 coding genes. The Venn diagram in Figure 4 illustrates the genetic variations and distinct characteristics that differentiate the genomes of HL26-1 and LL69-1 from those of closely related species. The protein-coding sequences of HL26-1 and LL69-1, along with those of three other Rhodotorula species (R. kratochvilovae CBS 7436, R. mucilaginosa GDMCC2.30, and R. paludigena P4R5), were compared to examine the similarity of their protein sequences. This comparison revealed that R. mucilaginosa HL26-1, R. kratochvilovae CBS 7436, R. mucilaginosa GDMCC2.30, and R. paludigena P4R5 possess 15, 84, 13, and 89 proteins, respectively, which are exclusive to their respective species or strains. Additionally, 11, 85, 35, and 36 protein families were identified as being species- or strain-specific for R. paludigena LL69-1, R. kratochvilovae CBS 7436^T, R. mucilaginosa GDMCC2.30, and R. paludigena P4R, respectively.

Figure 4. Venn diagram displaying the presence of conserved and specific proteins among Rhodotorula mucilaginosa HL26-1 (a) or R. paludigena LL69-1 (b) and R. kratochvilovae CBS 7436, R. mucilaginosa GDMCC2.30, and R. paludigena P4R5.

KEGG metabolic pathway analysis categorized a total of 2,795 (52.1%) protein-coding genes of HL26-1 into four major groups: metabolism (951 genes), genetic information processing (682 genes), cellular activities (225 genes), and environmental information processing (137 genes). Within the LL69-1 dataset, 3,303 protein-coding genes (48.7%) were subjected to KEGG analysis and classified into four major categories: metabolism (1,064 genes), genetic information processing (734 genes), cellular activities (245 genes), and environmental information processing (168 genes).

The putative gene families associated with astaxanthin biosynthesis in strains HL26-1 and LL69-1 were determined ( Table 2). The pathway of astaxanthin synthesis was divided into two main sections: one that creates the basic building blocks called terpenoids, which are important natural substances, and another that specifically produces astaxanthin.

The key genes associated with terpenoid biosynthesis discovered in the genomes of both strains were ACAT (acetyl-CoA acetyltransferase), HMGCS (hydroxymethylglutaryl-CoA synthase), HMGCR (hydroxymethylglutaryl-CoA reductase), MVD (mevalonate kinase), PMK (phosphomevalonate kinase), IDI (isopentenyl-diphosphate delta isomerase), GGDPS (geranylgeranyl diphosphate synthase), FDPS (farnesyl diphosphate synthase), and CrtE (geranylgeranyl pyrophosphate synthase). The diphosphomevalonate decarboxylase (DPMDC) gene has only been found in strain LL69-1. The multiple gene involved in astaxanthin biosynthesis were annotated in the genomes of both strains, including the CrtYB (phytoene synthase/lycopene beta-cyclase), CrtI (phytoene desaturase), CrtS (beta-carotene 4-ketolase/3-hydroxylase), and CrtR (cytochrome P450) genes.