The microorganisms and plasmids used or created in this work are listed in Table 1. The CBS 6938 strain of X. dendrorhous was used as the wild-type strain, and E. coli DH5α was used as a host for plasmid propagation. X. dendrorhous strains were cultured at 22°C with constant agitation at 170 rpm in YM medium (0.3% yeast extract, 0.3% malt extract and 0.5% peptone) supplemented with 1% glucose. Yeast transformant selection was performed on YM-agar 1.5% plates supplemented with 20 μg/ml hygromycin B and/or 35 μg/ml zeocin (Sigma, Saint Louis, MI, United States). E. coli strains were cultured with constant agitation at 37°C in LB medium (1% tryptone, 0.5% yeast extract and 0.5% NaCl) supplemented with 100 μg/ml ampicillin to select strains carrying plasmids and 40 μl of a 2% solution of X-gal (5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside) for recombinant clone selection (Sambrook and Russell, 2001, #31893). The recombinant clones carrying plasmids with the hygromycin B or zeocin resistance cassette were identified by direct colony PCR with specific primers for each cassette.

The ERG3 and ERG4 genes of X. dendrorhous were identified by local tBLASTn search over the genomic and transcriptomic databases from the X. dendrorhous strain UCD 67-385 (Baeza et al., 2015) with Geneious R11 using related sequences obtained from the GenBank database as queries. The exon-intron structure of each gene was manually cured by comparing genomic and transcriptomic data. Both gene sequences were uploaded to the GenBank database under the following accession numbers MN930922 and MN930923 for ERG3 and ERG4, respectively. The analysis of the promoter regions of the HMGS, ERG3 and ERG4 genes was carried out with the JASPAR¹ platform using the TRANSFAC database (Mathelier et al., 2016). The transcriptional factor SREBF2 was used as a consensus binding site. Protein sequence analyses were performed with programs available at http://www.ebi.ac.uk/interpro (Jones et al., 2014), https://embnet.vital-it.ch/software/TMPRED_form.html (Hofmann and Stoffel, 1993), and https://psort.hgc.jp/form2.html (Nakai and Horton, 1999).

Plasmid DNA was extracted from E. coli using the GeneJET Plasmid Miniprep Kit (Thermo Fisher, Waltham, MA, United States), and DNA sequencing was performed at Macrogen, Inc (Seoul, South Korea). The extraction of genomic DNA and total RNA from yeast X. dendrorhous was performed by mechanical rupture of a cell pellet obtained from a culture in liquid medium as described previously (Gómez et al., 2020, #70812). The obtained genomic DNA and RNA were stored at −20 and −80°C, respectively, until use.

All oligonucleotides designed and used in this work were synthesized by Integrated DNA Technologies (Coralville, IA, United States) and are listed in Supplementary Table S1.

To amplify the upstream and downstream fragments of the ERG3 and ERG4 genes, PCR analyses were performed using Pfu DNA polymerase (Agilent Technologies, Santa Clara, CA, United States). For E. coli recombinant clone selection and X. dendrorhous transformant confirmation, Taq DNA polymerase (Centro de Biotecnología, Facultad de Ciencias, Universidad de Chile, Chile) was used. PCR reactions were performed in a final volume of 25 μl using an Applied Biosystem 2720 thermocycler with the following program for standard PCR: initial denaturation at 95°C for 3 min (or 5 min in case of colony PCR or amplification with Pfu DNA polymerase), 35 cycles of denaturation at 94°C for 30 s, alignment of primers at 55°C for 30 s, and elongation at 72°C for 3 to 4 min followed by a final elongation step at 72°C for 10 min, and finally, the reaction was maintained at 4°C until analysis was performed.

Overlap extension PCR (OE-PCR) (Urban et al., 1997) was used to join the upstream and downstream DNA fragments of both studied genes, and the reaction was performed in a final volume of 25 μl with 1 U of Pfu DNA polymerase, PCR buffer 1X, each dNTP at 200 μM and 100 ng of each DNA fragment. The following program was used: initial denaturation at 94°C for 1 min, then 10 cycles of denaturation at 94°C for 30 s, DNA alignment at 55°C for 45 s, elongation at 72°C for 90 s, then a final elongation step at 72°C for 10 min, and finally the reaction was kept at 4°C until the next step. Then, the obtained reaction mixture was amplified by a standard PCR analysis with Pfu DNA polymerase and the same forward and reverse primers used to amplify the upstream and downstream DNA fragments, respectively.

RT reactions were performed in a final volume of 20 μl. First, 5 μg of total RNA in 11 μl of sterile water was mixed with 1 μl of oligo-dT (25 μM) and 1 μl of dNTPs (at 10 mM each) and incubated at 65°C for 5 min. Then, 1 μl of the enzyme Maxima Reverse Transcriptase (Thermo Fischer, Waltham, MA, United States), 4 μl of RT buffer 5X and 2 μl of DTT (0.1 M) were added. The mixture was incubated at 37°C for 52 min and then left at 70°C for 15 min, enabling it to cool at 4°C.

Real-time qPCR reactions were performed in a Mx3000P real-time PCR system (Agilent, Santa Clara, CA, United States) using Fast EvaGreen qPCR Master Mix (Bio-Rad Laboratories, Hercules, CA, United States). Samples were prepared in a final volume of 20 μl containing 1 μl of cDNA, 0.25 μM of each primer and 10 μl of kit reaction mixture. The efficiency of the primers used was greater than 95%, as determined by standard curves with a correlation coefficient R² ≥ 0.99. The obtained Ct (threshold cycle) values were normalized with respect to the corresponding value of the X. dendrorhous β-actin gene [GenBank: X89898.1] and expressed as fold change with respect to the control (Livak and Schmittgen, 2001).

The ERG3 and ERG4 genes were replaced by an antibiotic resistance cassette in the X. dendrorhous genome by homologous recombination using plasmids pΔErg3_Hyg and pΔErg4_Zeo (Table 1). Plasmids were constructed by joining the upstream (“UP”) and downstream (“DOWN”) regions of each gene of interest by OE-PCR, introducing an HpaI recognition site between them, according to the primer design (primers 5, 6, 15, and 16, Supplementary Table S1). Both regions were PCR-amplified from genomic DNA from X. dendrorhous strain CBS 6938, and the OE-PCR product was inserted at the EcoRV site of plasmid pBluescript SK to obtain plasmids pudErg3 and pudErg4. Then, each plasmid was digested with HpaI to insert either a hygromycin B or zeocin resistance cassette, which were obtained by EcoRV digestion of plasmids pMN-hph or pIR-zeo (Table 1), respectively.

Xanthophyllomyces dendrorhous was transformed by electroporation using a GenePulser Xcell^TM (BioRad, Hercules, CA, United States). Electrocompetent cells were prepared from cultures in YM medium at the exponential phase of growth at an OD_{600 nm} of 4–5 (Adrio and Veiga, 1995). Electroporation was performed using 6 μl of transforming DNA (10 μg) employing a 2 mm cuvette and the following conditions: 450 V, 125 μF and 600 Ω. After the pulse, 1 ml of YM was added to the cells, which were incubated for 4 h at 22°C before seeding in YM-agar 1.5% plates with the respective antibiotics to select the transformants.

Phenotypic analysis was performed using at least three biological replicas of each strain, including the mutant strains CBS.Δerg3 and CBS.Δerg4 and strains CBS.cyp61- and CBS 6938 as controls. Strains were cultivated in 100 ml of YM in 250 ml Erlenmeyer flasks at 22°C with constant agitation. Yeast growth was evaluated by the OD_{600 nm} of the culture, which was measured in a UV-vis V-630 spectrophotometer from JASCO. After 120 h of culture (stationary phase of growth), samples were taken to extract carotenoids, ergosterol and total RNA. This time point was selected as major differences in the transcript levels of gene HMGR of the MVA pathway, which is a known Sre1 gene target in X. dendrorhous (Gutiérrez et al., 2019), were observed when strain CBS.cyp61^– was compared to the wild-type (Loto et al., 2012).

Extracted carotenoids and sterols were quantified spectrophotometrically and normalized to the dry weight of the yeast. Carotenoids were extracted from cell pellets from 20 ml culture samples disrupted using glass beads and acetone (An et al., 1989) and quantified at 465 nm using an absorption coefficient of A_1% = 2,100. Sterols were extracted according to Shang et al. (2006). In brief, cell pellets from 5 ml culture samples were saponified in 16 ml of 60% ethanol and 4 g of KOH at 80°C for 2 h. Then, after the mixture was cooled, 5 ml of petroleum ether was added, vortexed for 10 s and centrifuged for 5 min. The supernatant was recovered, and this last step was repeated. Total sterols were quantified at 280 nm using an absorption coefficient of A_1% = 11,500.

Carotenoid and sterol compositions were evaluated by reverse-phase liquid chromatography (RP-HPLC) using a LiChrospher RP18 125-4 C-18 column (Merck Millipore, Billerica, MA United States). Carotenoid and sterol samples were run using acetonitrile: methanol: isopropanol (85: 10: 5, v/v/v) and methanol: water (97: 3, v/v), respectively, as the mobile phase with a 1 ml/min flux under isocratic conditions, and the elution spectra were recovered using a diode array detector. Sterols and carotenoids were visualized at the 280 and 474 nm channels, respectively, and compared to standards according to retention time and absorption spectra. The ergosterol standard was acquired from Sigma (Saint Louis, MI, United States).

Sterol identification was performed by a 1100 HPLC Series system equipped with a G1322A degasser and G1311A binary pump. For separation, a reversed-phased Zorbax SB-C18 analytical column (100 mm × 3.0 mm i.d., 5 μm particles) fitted with precolumn Zorbax SB-C18 was used. The mobile phase was prepared from methanol and acetonitrile 30: 70 (v/v). The flow rate was 0.6 ml/min, and 20 μl was injected. All solvents used were filtered through 0.5 ml Sartorius filters and degassed with ultrasounds. The flow rate was set to 0.5 ml/min, and the column temperature was set at 40°C. The mass spectra were acquired in positive mode using electrospray ionization, and analysis of all analytes was carried out in MRM mode. The other operating parameters were as follows: nebulizer gas flow, 3 L/min; drying gas flow, 15 L/min; desolvation line (DL) temperature, 250°C; and heat block temperature, 400°C. All chromatographic data were processed using ChemStation (v A09.03) software and Data Analysis (v 5.3).

The ERG3 and ERG4 genes from X. dendrorhous were identified by BLAST search over a genomic and transcriptomic databases from the yeast using homologous sequences from other organisms as queries. Once genes were identified, the exon-intron structure of each gene was manually cured by comparing the genomic and transcriptomic data. The ERG3 gene contains 6 exons of 554, 68, 102, 44, 105, and 114 bp, and the ERG4 gene has 15 exons of 202, 107, 97, 276, 134, 133, 4, 93, 35, 151, 16, 17, 55, 103, and 35 bp. From the coding sequence, it was deduced that the ERG3 gene encodes a polypeptide of 328 amino acids with an estimated molecular weight of 38.1 kDa and an isoelectric point (PI) of 7.88. In addition, the ERG3 polypeptide has 26 positively charged and 25 negatively charged residues. The ERG4 gene encodes a polypeptide of 485 amino acids. This polypeptide is different from the one deposited in the NCBI database, given that our CDS prediction contemplates the CDS of sterol reductase/beta-sheet receptor (CED83197.1) and C-24 sterol reductase (CED83198.1), with an estimated molecular weight of 55.1 kDa and a PI of 9.14. It has 31 negatively charged residues and 43 positively charged residues.

The subcellular localization of each protein was predicted using the PSORT II prediction tool. Both proteins are predicted to associate with the endoplasmic reticulum with 34.8% for Erg3p in the K-nn test and 55.6% for Erg4p, which is consistent with their potential function in the biosynthesis of ergosterol at the endoplasmic reticulum (Nishino, 1981). The number of possible transmembrane segments in each of them was also predicted by using the TMHMM Server v. 2.0, where it was observed that Erg3p has 3 transmembrane segments (Figure 2A), while Erg4p has 9 (Figure 2B). Additionally, through the InterPro platform, the presence of conserved domains for both proteins were predicted in silico. Erg3p contains 1 domain of the fatty acid hydroxylase superfamily (IPR006694), while for Erg4p, two conserved domains belonging to the family of sterol reductase proteins were observed (IPR018083).

To evaluate the functionality of both identified genes, they were independently replaced by an antibiotic resistance module in the X. dendrorhous genome. To this end, recombination modules were constructed and inserted at the EcoRV site of plasmid pBluescript II SK- to obtain plasmids pCBS-Δerg3 and pCBS-Δerg4, respectively (Table 1). The ERG3 recombination module contained 585 bp upstream and 538 bp downstream of the gene flanking a hygromycin B resistance cassette, while the ERG4 recombination module contained 610 bp upstream and 579 bp downstream of the gene and a zeocin resistance cassette between them.

Once plasmids pCBS-Δerg3 and pCBS-Δerg4 were obtained, the wild-type strain CBS 6938 was transformed to replace the respective gene by the antibiotic resistance module through homologous recombination. In this way, three transformants resistant to hygromycin B and four transformants resistant to zeocin were obtained for the ERG3 and ERG4 gene mutations, respectively. To confirm the gene replacement events on the obtained transformants, their genomic DNA was extracted, and PCR analyses were performed using an informative set of primers (Figure 3). For the following analysis, a mutant of each gene was randomly selected.

Four strains were analyzed and compared: the wild-type strain CBS 6938 and the mutant strains CBS.Δerg3, CBS.Δerg4 and CBS.cyp61^– (Figure 4A). Only strain CBS.Δerg4 showed a different growth curve, reaching a lower OD at the stationary phase of growth (Supplementary Figure S1).

Considering that the genes ERG3 and ERG4 are involved in the biosynthesis of ergosterol, total sterols were extracted and quantified (Figure 4B). A significant difference was observed in the total amount of sterols produced by the mutant strains CBS.Δerg3 and CBS.Δerg4 with respect to the parental strain CBS 6938; both mutants have a higher sterol content than the wild-type strain. Neither strain CBS.Δerg3 nor CBS.Δerg4 produced ergosterol but rather accumulated other sterols with different retention times (Supplementary Table S2). In addition, to determine the complete profile of sterols in each strain, an LC-MS/MS analysis was performed (Table 2). It was observed that the major compound in strain CBS 6938 corresponds to ergosterol, and it also accumulates other sterols from different levels of the ergosterol synthesis pathway. None of the mutant strains analyzed in this work accumulated ergosterol, and they had a different sterol profile according to each mutation (the affected enzyme). Thus, the main intermediary accumulated in strain CBS.Δerg3 corresponded to episterol. In this case, the affected enzyme converts episterol into ergosta-5,7,24(28)-trienol. The strain CBS.Δerg4 accumulated ergosta-5,7,22,24(28)-tetraenol since the affected enzyme catalyzes the conversion of this molecule to ergosterol. Finally, the CBS.cyp61^– strain accumulated ergosta-5,7,24(28)-trienol and episterol as major compounds because the affected enzyme catalyzes the conversion of ergosta-5,7,24(28)-trienol to ergosta-5,7,22,24(28)-tetraenol. In addition, all mutants accumulated upstream sterols, such as lanosterol, zymosterol and fecosterol. In summary, this analysis confirmed that none of the analyzed mutants accumulated ergosterol.

Previously, a different color phenotype was observed in cyp61^– mutants with respect to their parental wild-type strains (Loto et al., 2012); however, this difference was not observed in the CBS.Δerg3 or the CBS.Δerg4 mutants. To confirm these observations, carotenoids from the analyzed strains were extracted and quantified: only the CBS.cyp61^– mutant had a significantly higher carotenoid content than the other strains (Figure 4C). Carotenoid composition was also analyzed, and minimal differences were observed among strains (Table 3). These results indicate that even though the CBS.Δerg3 and CBS.Δerg4 mutants do not produce ergosterol, carotenoid production is not enhanced, as was observed in the X. dendrorhous cyp61^– mutants, suggesting that is not the absence of ergosterol what enhances carotenoid production but probably the different sterol composition in this mutant would favor carotenoid production.

In a previous work, we observed that cyp61^– mutants derived from different X. dendrorhous wild-type strains had higher transcript levels of the HMGR gene, which is involved in the MVA pathway, in relation to the corresponding parental strains (Loto et al., 2012). Transcriptional activation of several genes involved in the MVA pathway and in the synthesis of ergosterol in different organisms is regulated by sterol levels through proteins SREBPs (named Sre1 in several fungi, including X. dendrorhous). In general, at low sterol levels, SREBP/Sre1 is activated by proteolytic cleavage, allowing the transcription factor domain to travel to the nucleus and bind to specific DNA sequences, known as sterol regulatory elements (SREs), in the promoter region of target genes to regulate their expression. Conversely, when sterol levels increase, SREBP/Sre1 cleavage is inhibited; therefore, the synthesis of sterols is reduced through a negative feedback mechanism (Hughes et al., 2005). Considering that the ERG3 and ERG4 genes encode enzymes involved in ergosterol biosynthesis such that they could be regulated by Sre1, the promoter region of these genes was analyzed to search for potential SREBP/Sre1 binding sites. For comparative purposes, the promoter region of the HMGS gene was also analyzed, as it was demonstrated that Sre1 in X. dendrorhous binds to this DNA (Gutiérrez et al., 2019). The consensus SRE sequence recognized by the human transcription factor SREBF2 was predicted in the promoter region of the three analyzed genes: HMGS has four potential SRE sequences: one located at the [+] strand and three at the [−] strand, ERG3 has two potential SREs at the [+] strand and ERG4 has one potential SRE at the [−] strand (Supplementary Table S3).

To obtain further insight, the relative transcript levels of genes ERG3, ERG4 and HMGS were evaluated by RT-qPCR after 120 h of culture in the four X. dendrorhous strains studied in this work (Figure 5). Interestingly, transcript levels of the three genes were higher in strain CBS.cyp61^–. Considering this result and that potential SRE sequences were identified in their promoter regions, their expression at the transcriptional level could indeed be regulated by Sre1, which is in its active form in strain CBS.cyp61^–. Although mutants CBS.Δerg3 and CBS.Δerg4 do not produce ergosterol, in general, the transcript levels of the three genes did not change with respect to the wild-type strain. The only exception was the HMGS gene in strain CBS.Δerg4, which showed higher transcript levels than the wild-type strain; however, this increment was only approximately 2-fold in contrast to 20-fold observed in strain CBS.cyp61^–. In addition, the ERG3 and ERG4 transcripts were not detected in the respective mutant strains, confirming that these mutants do not express these genes.