Precultures of D. hansenii wild-type strain CBS767 (WT) and the isogenic mutant Dhhog1Δ, generated by Sánchez et al. (2020) were grown and maintained on Yeast Peptone Dextrose agar (YPD agar; 1% yeast extract, 2% peptone, 2% glucose, and 2% agar). Liquid precultures were prepared in YPD medium using a flask-to-medium ratio of 2:5, incubated at 28°C with shaking at 180 rpm overnight, representing the seed condition (time 0 h). Cells were recovered by centrifugation at 3,000 rpm for 5 min, washed twice, and resuspended in sterile distilled water. Subsequently, cultures were grown in minimum media based on Yeast Nitrogen Base (YNB) supplemented with 2% glucose, 0.5% ammonium sulfate as the sole nitrogen source, and 0.6 M NaCl (from now on referred to as minimum media + 0.6 M NaCl).

Washed preculture cells were used as an inoculum for the minimum medium at an initial optical density of 0.05 (measured at 600 nm, OD_{600 nm}). Successive OD_{600 nm} measurements were taken using a Beckman Coulter DU^® 640 spectrophotometer. For the flavinogenic condition (minimum media + 0.6 M NaCl, initial pH 4.3), the exponential growth phase was defined at OD_{600 nm} = 0.5 (18–20 h), whereas the stationary growth phase was defined at OD_{600 nm} = 3.8–4.0 (48 h).

For pH experiments, the basal pH of the minimum media YNB (initially 4.3) was adjusted to 6.8–7.0 using 2 N NaOH. pH measurements were carried out at defined time intervals with a potentiometer (pH 210, Microprocessor pH Meter, Hanna Instruments), previously calibrated according to the manufacturer’s instructions.

Extracellular riboflavin was quantified by measuring fluorescence intensity at emission/excitation wavelengths of 440/535 nm. A riboflavin stock solution (20.1 mg/L) was prepared by dissolving riboflavin powder in distilled water under mild heating (<30°C). Culture aliquots (15 mL) were collected at multiple time points ranging from 24 to 122 h to capture the onset and dynamics of riboflavin production, as indicated by the fluorescence and yellow coloration of the supernatant, covering both exponential and stationary growth phases and the associated pH decrease. Aliquots were centrifuged, and supernatant was filtered through 0.22 μm pore-size filters. Subsequently, 200 μL of the filtered supernatants were transferred to black 96-well Costar^® assay plates with clear flat bottoms, and fluorescence was measured with a Synergy H1 Microplate Reader (Biotek), gain = 50. Riboflavin concentrations were determined by interpolation against a standard curve generated from 1:2 serial dilutions of an initial 10.05 mg/L riboflavin solution.

Sterile culture supernatants were lyophilized to concentrate the fluorescent compound and remove water. The resulting dried powder was extracted with 50 mL HPLC-grade methanol and vortexed for 5 min at room temperature (25 ± 1.0°C). Extracts were centrifuged at 3,500 rpm for 10 min, and the methanolic phase was recovered and dried at 35°C for 72 h. Reversed-phase high-performance liquid chromatography (RP-HPLC) coupled with a Diode Array Detector (DAD), was then carried out, following the method described by Jiménez-Nava et al. (2023). Methanolic residues were resuspended in 10 mL HPLC-grade water, and 2 mL aliquots were loaded in triplicate onto C18 solid-phase extraction (SPE) cartridges (Alltech^® Maxi-Clean™, Thermo Fisher Scientific, Waltham, MA, United States). Cartridges were eluted under vacuum with 1.2 mL of HCl-acidified methanol (0.002% v/v). Eluates were analyzed using an Agilent 1,260 Infinity Series system equipped with a DAD, monitoring 280 and 440 nm, with a Zorbax SB-Phenyl column (150 × 4.6 mm, 5 μm, Agilent Technologies, Santa Clara, CA, United States). The mobile phase was acetonitrile and 0.05% phosphoric acid in water at 0.5 mL/min, applying a linear gradient of 10–100% acetonitrile over 15 min.

Absorption spectra of the HPLC-separated fractions were obtained directly from DAD data and visualized with Agilent ChemStation software, using a riboflavin standard (Sigma-Aldrich, Cat. No. 47861) for reference.

Culture supernatants (15 mL) were collected at 18 h and 48 h from independent biological cultures and clarified at room temperature by centrifugation (3,000 rpm, 5 min). Clarified supernatants were transferred to clean tubes and filtered through 0.22 μm membranes prior to analysis. No acid digestion was performed; therefore, measurements correspond to the dissolved fraction of the clarified supernatant. Elemental quantification was performed by Inductively Coupled Plasma–Optical Emission Spectroscopy (ICP-OES) by an external analytical service using the proprietary Triton ICP-OES Test (TRITON Applied Reef Bioscience, Germany).¹ Due to the proprietary nature of the service, detailed instrument parameters are not disclosed by the provider, and results are reported as a single quantitative concentration output per submitted sample. Elements quantified included B, Ba, Be, Br, Ca, Cl, Co, Cr, Cu, F, Fe, I, K, Li, Mg, Mn, Mo, Na, Ni, P, PO4, S, Si, Sr, V, and Zn (Supplementary Table 1). Element concentrations are reported in mg/L or μg/L and converted to μM (Supplementary Table 1). Detection limits were those provided by the facility, and all reported values were above the sensitivity threshold.² All samples were analyzed under identical matrix conditions (same culture medium, pH, and NaCl concentration), enabling direct comparisons between strains and time points.

Total RNA was extracted following the protocol described by Schmitt et al. (1990). Cultures were first grown overnight in rich medium (YPD) to obtain seed cultures (time 0 h), and subsequently inoculated into minimum media with 0.6 M NaCl. Cells were harvested at exponential phase and stationary phase.

Briefly, cells were washed twice, centrifuged at 3,000 rpm, and mechanically disrupted using a vortex mixer with sterile glass microbeads (425–600 μm) previously incubated in phenol (pH 4.5). Samples were incubated at 65°C for 5 min and vortexed twice for 30 s. The mixture was chilled and centrifuged to separate the aqueous and phenolic phases. The aqueous phase was extracted twice with phenol:chloroform:isoamyl alcohol (25:24:1) and chloroform:isoamyl alcohol (24:1). RNA was precipitated by adding 1/10 volumes of sodium acetate (3 M, pH 5.2) and 2.5 volumes of absolute ethanol, followed by incubation at −20°C for 30 min and centrifugation at 14,000 rpm. The resulting pellet was washed, dried, and resuspended in RNase-free water. RNA integrity was assessed by electrophoresis on a 1% denaturing agarose gel.

The D. hansenii genome remains only partially annotated, with most gene assignments based on in silico predictions of orthologous genes. To validate these annotations, multiple sequence alignments of selected riboflavin biosynthesis proteins and the transcription factor Sef1 were carried out against their orthologous in Saccharomyces cerevisiae and Candida albicans (Supplementary Material 1). Amino acid sequences were further compared to confirm that the predicted ORFs corresponded to the expected protein functions (RIB1, locus DEHA2A12870g; RIB2, locus DEHA2E11374g; RIB4, locus DEHA2D04180g; RIB5, locus DEHA2D13926g; RIB6, locus DEHA2G09504g; RIB7, locus DEHA2G10010g; SEF1, locus DEHA2C16676g). Protein identifiers (IDs) were retrieved from the NCBI Protein Database. Both analyses were performed within the NCBI BlastP Tool.

The intergenic region spanning from −625 bp upstream to the start codon of each gene was scanned with position-specific scoring matrices (PSSMs) using the Regulatory Sequence Analysis Tools (RSAT) platform,³ employing the full matrix-scan tool within the Fungi Server (Turatsinze et al., 2008). Binding motifs for Hog1-regulated transcription factors were identified, including Sef1 (ID: 2900038, locus DEHA2C16676g), Hot1 (ID: 2901482, locus DEHA2D15136g), Sko1 (ID: 2901375, locus DEHA2D09196g), Skn7 (ID: 2913769, locus DEHA2B08052g), Msn2/4 (ID: 2899670, locus DEHA2A08382g), and Yap1 (ID: 2904514, locus DEHA2G02420g).

Except for Sef1 and Hot1, motif matrices were retrieved from the JASPAR database, specifically from the 2024 core fungi collection: Sko1 (ID: MA0382.1), Skn7 (ID: MA0381.1), Msn2/4 (ID: MA0341.1/MA0342.1), and Yap1 (ID: MA0415.1). A D. hansenii-specific background model estimation method with a Markov order of 1 was applied, and a significance threshold of p-values < 0.0006 was used for significant determination. Putative Sef1 binding sites were searched manually in each intergenic region based on the DNA recognition motifs identified by Chen et al. (2011) and Romanov et al. (2025) for C. albicans and C. famata Sef1, respectively. Two distinct putative Hot1 binding sites were searched manually based on the motifs reported by Cook and O’Shea (2012), Bai et al. (2015) and Gomar-Alba et al. (2015) for S. cerevisiae (Supplementary Figure 2). The Sef1 consensus binding motif logo was derived from multiple alignments of the upstream intergenic regions of RIB genes with reported sequences from C. albicans and C. famata (Chen et al., 2011; Romanov et al., 2025; Supplementary Figures 3, 4).

To preserve RNA integrity and minimize RNase contamination during downstream gene expression analyses, all RNA handling steps were performed under RNase-free conditions using RNase-free plasticware and reagents. RNA integrity was assessed by agarose gel electrophoresis prior to reverse transcription. For RT-qPCR analyses, total RNA samples were treated with DNase I to remove residual genomic DNA before cDNA synthesis. First-strand cDNA was synthesized using the Thermo Scientific RevertAid H Minus Kit (K1632), which includes the RiboLock RNase inhibitor to protect RNA during reverse transcription. The resulting cDNA was subsequently used as a template for quantitative PCR (qPCR).

RT-qPCR assay was performed using the standard curve method with gene-specific deoxyoligonucleotides designed for the genes encoding Hog1 MAPK (DhHOG1), sugar transporter-like protein (DhSTL1), the transcriptional activator Sef1 (DhSEF1), GTP cyclohydrolase II (DhRIB1), DArPP deaminase (DhRIB2), 6,7-dimethyl-8-ribityllumazine synthase (DhRIB4), riboflavin synthase (DhRIB5), 4-dihydroxy-2-butanone-4-phosphate synthase (DhRIB6), 5-amino-6-(5-phosphoribosyl-amino) uracil reductase (DhRIB7). The actin gene (DhACT1) was used as the housekeeping control.

Deoxyoligonucleotides were evaluated to ensure the absence of dimer formation and cross-hybridization, and only deoxyoligonucleotides pairs with amplification efficiencies > 90% were used (Table 1). RT-qPCR was performed with a Rotor-Gene Q thermocycler (Qiagen) using SYBR Green as the detection dye (KAPA SYBR Fast Kit, Roche). Cycling conditions were: 94°C for 5 min (1 cycle), followed by 35 cycles of 94°C for 15 s, 60°C for 20 s, and 72°C for 20 s.

Relative transcript levels were normalized to the WT mean Ct for each gene using the 2^∧−ΔΔCt method, and data were expressed as fold change values. Results represent mean ± SD from three biological replicates, each analyzed in duplicate (technical replicates).

Unpaired t-tests were applied to compare the means of the WT and Dhhog1Δ groups within each time point or gene in GraphPad Prism (GraphPad Software Inc.). Optical density, pH, riboflavin quantification and relative expression means were based on three independent experiments (n = 3). For riboflavin quantification and gene expression, statistical significance was indicated as follows: p ≤ 0.05 (*), p ≤ 0.005 (**), p ≤ 0.0005 (***), and p ≤ 0.00005 (****). Results are presented as the mean of three independent measurements ± standard deviation (SD).

For in silico predictions of transcription factor binding sites (TFBS), a D. hansenii-specific background model estimation method with a Markov order of 1 was applied, and a significance threshold of p-values < 0.0006 was used for the determination of putative binding sites in select riboflavin-related genes.

The high-osmolarity glycerol (HOG) pathway, with the MAP kinase Hog1 as its terminal effector under salt stress, is recognized as the most conserved signaling cascade for stress adaptation in fungi (de Nadal et al., 2002; Hohmann, 2002; Hohmann et al., 2007; Hohmann, 2009; Saito and Posas, 2012). Given previous observations of riboflavin efflux in acidic media (Perl et al., 1976; Vanetti and Aquarone, 1992), we investigated whether the initial medium pH modulates growth and riboflavin excretion in D. hansenii under saline stress, raising the possibility that DhHog1 signaling may also influence riboflavin metabolism. Comparative assays were conducted between the WT strain and the Dhhog1Δ mutant in minimum media (YNB) with glucose (2%) supplemented with 0.6 M NaCl under either neutral or acidic initial pH conditions. Significant differences in biomass and cell viability were observed during the stationary phase (Figures 2A,B), highlighting the role of DhHog1 in maintaining cellular fitness under osmotic stress.

Remarkably, cultures initiated at low pH exhibited secretion of a yellow pigment as early as the stationary phase in the Dhhog1Δ mutant (48 h) (Figure 2C), whereas in the WT strain, pigment accumulation became evident only at the late stationary phase (122–172 h) (Figure 3A). By contrast, no pigment secretion was detected in either strain when cultures were initiated at neutral pH (Figure 2C), even after 5 days of growth. Thus, NaCl was not sufficient to trigger pigment secretion at neutral pH, supporting a cooperative effect between acidic pH and salinity. These findings indicate that initial acidic conditions are a key determinant for pigment production and excretion in D. hansenii, with an accelerated onset in the absence of HOG1.

Previous studies have reported that D. hansenii is capable of producing riboflavin, a metabolite responsible for the characteristic yellow coloration (Gadd and Edwards, 1986; Seda-Miró et al., 2007). Based on this evidence, we hypothesized that the yellow pigment observed under our culture conditions could correspond to riboflavin. To test this, we identify the pigment by RP-HPLC-DAD analyses in the culture supernatants from the Dhhog1Δ mutant, revealing the presence of a major fluorescent compound. Retention times closely matched those of a riboflavin standard (8.450 ± 0.003 min). The Dhhog1Δ mutant exhibited clearer chromatographic separation and UV-Vis absorption peaks characteristic of flavins (λmax = 223, 268, 370, 445 nm), consistent with riboflavin identity. In contrast, WT supernatants displayed weaker signals with partially overlapping absorption peaks. Representative chromatograms and UV-Vis spectra are shown in Figures 4A–F.

Furthermore, we examined the fluorescence properties of the riboflavin by recording excitation and emission spectra at the characteristic wavelengths (excitation: 440 nm; emission: 520–535 nm). Extracellular riboflavin production was monitored over time (24–172 h) (Figure 3A) to capture its onset and dynamics, as indicated by fluorescence and the characteristic yellow coloration of the supernatant. Sampling covered the exponential and stationary growth phases. Riboflavin concentrations were quantified by fluorimetry using a standard calibration curve (Figures 3B–D). These experiments confirm that the accumulation of riboflavin in the supernatant of cultures initiated at low pH, accumulates more rapidly over time in the Dhhog1Δ mutant. This observation indicates, for the first time, that DhHog1 is involved in the timing of riboflavin accumulation in D. hansenii under acidic and saline conditions.

Iron limitation and the availability of other essential elements have been reported as key triggers for riboflavin excretion in riboflavin-producing yeasts (Averianova et al., 2020; Ruchala et al., 2025). To determine whether differences in elemental uptake contribute to the enhanced riboflavin accumulation observed in the Dhhog1Δ mutant, culture supernatants from 18 and 48 h under saline conditions and initial low pH were analyzed by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) (Figure 5). Absolute concentrations (mg/L) and molar conversions (μM) are provided in Supplementary Table 1. Elemental profiles are interpreted as quantitative trends supporting comparisons between strains and time points.

By 48 h, the Dhhog1Δ mutant displayed more rapid assimilation of several essential elements, including phosphorus, phosphates, sulfur, and magnesium, compared to the WT strain. Notably, the most pronounced differences were observed for phosphorus/phosphate (1.75E+06/1.74E+06 μM), sulfur (3.22E+04 μM), and magnesium (2.14E+03 μM), consistent with accelerated assimilation of macronutrients in the Dhhog1Δ strain (1.10E+06/1.09E+06, 2.76E+04, and 1.65E+03 μM) during the transition to stationary phase (Figure 5). In contrast, uptake of iron and other elements was similar in both strains (Supplementary Table 1), indicating that mechanisms other than iron uptake are likely responsible for stimulating riboflavin excretion in D. hansenii. These results suggest that accelerated assimilation of specific elements (P, S, and Mg) in the Dhhog1Δ mutant may contribute to the earlier and higher riboflavin production observed under acidic and saline conditions.

The riboflavin biosynthesis genes RIB1, RIB2, RIB5, RIB6, and RIB7 were previously cloned from the synonymous species Candida famata (Voronovsky et al., 2002, 2004; Dmytruk et al., 2004), and their sequences are available in NCBI GenBank as part of the Genolevures Consortium’s annotated reference genome for D. hansenii (ASM644v2). The RIB4 gene sequence was inferred from its homologs in S. cerevisiae and C. albicans and incorporated into the Genolevures Consortium reference genome (Morgunova et al., 2007; Skrzypek et al., 2017). Amino acid sequence alignments confirmed the identity of the main riboflavin biosynthetic enzymes and revealed the degree of sequence conservation among them (Table 2).

To investigate the potential regulation of riboflavin biosynthesis genes by DhHog1 in D. hansenii, we performed an in silico analysis to identify putative stress-related transcription factor (TF) binding sites within the promoters of RIB genes and the transcription factor Sef1 (Figures 6A,B). Initially, a motif analysis was performed, and sequence logos were generated to represent putative SEF1 binding motifs derived from multiple alignments of the upstream intergenic regions of RIB genes (Figure 6A) with previously reported sequences from Candida albicans and C. famata (Chen et al., 2011; Romanov et al., 2025). For the analysis of additional motifs, we relied on the previous work by de la Fuente-Colmenares et al. (2024), which validated the likely conservation of function of Sko1, Skn7, Msn2/4, and Yap1 proteins in D. hansenii through in silico comparisons with their homologs in S. cerevisiae and C. albicans (de la Fuente-Colmenares et al., 2024).

Our in silico intergenic region analysis revealed that the RIB gene promoters contain putative binding motifs for Hog1-controlled transcription factors Hot1, Sko1, Msn2/4, Yap1, and Skn7, as well as for the iron metabolism regulator Sef1 (Figure 6B). Specifically, Sef1-binding sequences were found predominantly in the promoters of RIB1, RIB2, RIB4, and RIB6. The two reported variants of Hot1-binding sequences in S. cerevisiae (Cook and O’Shea, 2012; Bai et al., 2015; Gomar-Alba et al., 2015) were only found in the RIB4 promoter at a distance of 69 nt of each other. Sko1 motifs were present in RIB2, RIB5, and SEF1 promoters. Skn7-binding sequences were detected in RIB1, RIB2, RIB4, and RIB7. Msn2/4 motifs were found in RIB1 and RIB4, and Yap1-binding sites in RIB2 and RIB6.

The absence of HOG1 expression was confirmed in the Dhhog1Δ mutant, accompanied by reduced expression of STL1, a downstream HOG pathway gene involved in glycerol symport (Figure 6C). Expression profiles further reveal that, in the absence of DhHog1; SEF1, RIB1, RIB4, and RIB6 transcripts accumulate at higher levels during the logarithmic phase, with upregulation of RIB1, RIB4, and RIB6 genes persisting into the stationary phase in the Dhhog1Δ mutant (Figures 6D,E). Interestingly, RIB2 and RIB7 did not show increased expression despite riboflavin oversynthesis (Figures 6D,E). This suggests possible differential regulation of these genes, which may involve promoter-specific mechanisms or post-transcriptional control. These findings indicate that DhHog1 influences riboflavin biosynthesis in D. hansenii, possibly through the coordinated activity of stress-responsive transcription factors such as Hot1, Sko1, Skn7, Msn2/4, and/or Yap1, together with the iron metabolism regulator Sef1.