Reversible naftifine-induced carotenoid depigmentation in Rhodotorula mucilaginosa (A. Jörg.) F.C. Harrison causing onychomycosis

Rhodotorula mucilaginosa was isolated from a patient with onychomycosis, and identification was confirmed by morphological and cultural characteristics as well as by DNA molecular analysis. Antifungal agents naftifine (10 mg/mL, active substance in Exoderil) and bifonazole (10 mg/mL, active substance in Canespor) were tested in different concentrations to assess in vitro effects on fungal growth and carotenoid synthesis. The antifungal mechanisms of action of naftifine and bifonazole against R. mucilaginosa isolates were similar and affected the biosynthetic pathway of ergosterol. For the first time, this research demonstrates that naftifine affects the carotenoid biosynthetic pathway, producing depigmentation of R. mucilaginosa in solid and liquid media. Furthermore, depigmentation was a reversible process; naftifine-treated yeast cells that were depigmented resumed carotenoid production upon transfer to fresh media. Raman and UV-vis spectrophotometry in conjunction with chromatographic analysis detected changes in carotenoids in yeast cells, with torulene decreasing and B-carotene increasing after repigmentation. Transmission electron micrographs revealed critical ultrastructural modifications in the depigmented cells after naftifine treatment, i.e., a low-electron-density cell wall without visible mucilage or lamellate structure.


Results and Discussion
The results indicated that R. mucilaginosa was a causative agent in this specific case of onychomycosis in an aged patient with chronic HBV hepatitis (Fig. 1A). After incubation of all toenail samples in SDA control media, R. mucilaginosa isolates were identified by morphological and cultural characteristics. Control colonies of R. mucilaginosa obtained on SDA were mucoid, with red colour (Fig. 1B) and a diameter of 12-13 mm 3 days after inoculation. Molecular analysis was performed for species identity confirmation 14,23 . The DNA sequence was run against the BLAST-NCBI nucleotide database 31 as a query and matched identically with R. mucilaginosa (GenBank: KU052792.1) species.
Then, by exposing a control R. mucilaginosa isolate for 10 minutes to 1 mg/mL naftifine solution, depigmented fungal colonies were obtained on SDA control media. The depigmented colonies were mucoid and had similar diameters to the red colonies but showed a milky colour. Moreover, by including small concentrations of naftifine (0.1 mg/mL, 0.2 mg/mL) in SDA before inoculation of R. mucilaginosa isolate, an inhibitory effect was obtained compared to the control colonies (Fig. 1C). These depigmented colonies had a diameter of 7-8 mm on SDA with 0.1 mg/mL naftifine and of 5-6 mm on SDA with 0.2 mg/mL naftifine, showing a small dose-dependent inhibitory effect of the drug (Fig. 1C). Furthermore, by inoculating depigmented R. mucilaginosa isolate into SDA control media, colonies regained pigment (Fig. 1D). Considering that carotenoids are important substances against reactive oxygen species (ROS), depigmentation may be correlated with increased oxidative stress and fungal injury. Repigmentation suggested that the carotenoid metabolism changes induced in the fungi by naftifine were reversible. These surprising results suggested further analysis regarding R. mucilaginosa carotenoid metabolism and related structural changes.
Preliminary tests of R. mucilaginosa growth under our conditions indicated that the stationary phase was reached after 60-70 h. However, in the initial exponential stage the cells appeared mostly colourless (low carotenoid content), and only after approximately 50 h could the pigmentation be observed. This monitoring was performed at various pH values (pH range 3.5-7.5), especially for tracking carotenoid formation. The growth was pH-dependent only in the initial exponential phase (Fig. S1A). Although the pH profile change in time was dependent on the initial pH (Fig. S1B), indicating a pH-dependent metabolism for R. mucilaginosa, carotenoid formation in terms of both rate and composition did not depend on pH significantly, as observed by monitoring the intensity and spectral features from the resonant Raman spectra (Fig. S2). We aimed to further understand the factors that influenced and controlled carotenoid formation in Rhodotorula spp. beyond the factors that interfered with basic metabolism, such as type of C and N sources and their ratio 26 , compounds inducing ROS stress (e.g., duroquinone, methylene blue). ROS upregulated carotenoid content in R. mucilaginosa, indicating that the carotenoids were a response of the cells to the negative effects of ROS agents (e.g., superoxide, singlet oxygen) as previously shown 28,32 . Other chemicals, such as amphotericin and diphenylamine, not only inhibited carotenoid formation but also cell development (Fig. S3). In addition, we hypothesized that the photosensitizers hypericin and hyperforin could also induce carotenoid formation; however, this was not the case (Fig. S3), most likely due to the low level of the induced potential stress.
During these initial experiments, we observed that one of the two tested antifungal agents, namely, naftifine (an allylamine antifungal drug), did not inhibit cell growth at lower concentrations, but rather induced depigmentation of the cells when growing on solid medium (Fig. 1C). The same results were obtained in liquid media, in controlled conditions. In addition to determination of naftifine MIC50 (55 ± 14 mg/L) for its antifungal activity ( Fig. 2A), quantified by OD 600nm, the effect of naftifine on depigmentation was assessed by OD 512nm and a corresponding MIC50 (0.088 ± 0.02 mg/L) was determined (Fig. 2B). Depigmentation did not affect the cell growth at these naftifine concentrations, as could be observed by the optical density values. In contrast, bifonazole (an imidazole antifungal drug) did not produce any depigmentation in R. mucilaginosa, although its MIC50 for antifungal activity was one order of magnitude lower, 4.4 ± 0.8 mg/L ( Fig. 2C and D). Similar depigmentation was observed by resonant Raman spectroscopy (Fig. 3A), where besides the expected decrease in signal intensities of the main peaks 999 cm −1 (C = CH deformation), 1150 cm −1 (symmetrical C-C stretching), 1505 cm −1 (symmetrical C = C stretching) due to carotenoid content, subtle changes could be observed in the 1200-1400 cm −1 region. Minor Raman shifts were recorded for the secondary peak at 1283 cm −1 (torulene), ascribed to the CH 2 deformation in carotenoids 33,34 . This band was shifted in the case of torularhodin at 1284 cm −1 , at 1286 cm −1 for β-carotene and at 1290 cm −1 for unidentified carotenoid (C nd) , indicating a possible change of the carotenoid composition in the cells, thus a qualitative change. This observation was further supported by the UV-vis spectra, where in addition to an absorbance decrease with increased naftifine concentrations, the spectral profiles also changed (i.e., the ratio of the three main spectral features) (Fig. 3B). Next, for more in-depth investigations, chromatographic analysis of the extracted carotenoids was performed. From a variety of solvents or conditions for cell disruption available in the literature, we chose simple DMSO treatment of the washed cells. The advantages of this procedure are the high yield of extraction (only a colourless pellet remained, indicating a quantitative extraction) and the good solubility of ergosterol in this solvent, which allowed a simultaneous extraction of both type of constituents. The High-Performance Liquid Chromatography (HPLC) analysis allowed the separation of four main carotenoids including torularhodin, torulene and β-carotene (C1, C2 and C3, respectively), in good agreement with previous data 27,32 (Fig. 4A), while the other compound (C nd ) could not be identified. The identification was done using a pure standard for β-carotene and retention times, as well as using highly specific spectral features of the other two compounds, which were in very good agreement with already known data (Fig. 4B, Table 1). The unidentified compound presented torulene-like carotenoid absorption spectral features and had a chromatographic behaviour (lower retention time allowing assignment as hydroxy-torulene) resembling previously detected carotenoids in other Rhodotorula spp. 35 . Additionally, ergosterol was also separated using the same method and identified based on its chromatographic and highly specific spectral characteristics (Fig. 4A). Moreover, using a dedicated method for ergosterol extraction, a highly similar spectrum was obtained. Therefore, all these related components could be quantitatively evaluated, allowing a comparative analysis for cells growing in different conditions. Thin-Layer Chromatography (TLC) was also used for carotenoid separation after DMSO removal using extraction cartridges. Using this method, the same three components were separated and identified, followed by their Raman detection directly on the plate. The contour map of the detected signal is presented in Fig. 4C, and the individual Raman signals are shown in Fig. 4D. As expected, the spectra had a high degree of similarity due to highly similar chemical structures (Fig. 4E); nevertheless, subtle changes could be observed at approximately 1280 cm −1 , and a slight spectral shift appeared at 1505 cm −1 band (Table 1, Fig. 4D). In addition to torularhodin, torulene and β-carotene, two unidentified carotenoids were also observed (probably different than the one from HPLC analysis, due to retention time analysis). This may have occurred because in the introduction of an extra sample preparation step (cartridge extraction for DMSO removal), analyte alteration could occur, especially for those present in lower quantities.
A Principal Component Analysis (PCA) was applied to the absorbance spectra of the DMSO extracts of the cells grown under various naftifine concentrations, and the plot of the scores corresponding to the first two principal components are shown in Fig. 5A. Surprisingly, two different clusters were noted, one with samples at low concentrations and one with samples at high concentrations, together with a transition zone. Such a grouping indicated different compositions of carotenoids with increasing naftifine concentration. To eliminate the concentration effect (decrease in intensity as the carotenoid content is diminished), for PCA analysis normalized spectra were used, thus allowing a qualitative-only analysis (relative content of carotenoids). The normalized spectra are presented in Fig. 5B. A shift of the spectral maxima towards the UV region was observed (the solutions are in DMSO, and thus the maxima were different from those in the HPLC data, where the solvent was different). This shifting of the maxima clearly indicates an increase of β-carotene relative content and decrease of torularhodin relative content as the naftifine concentration was increased. This was in very good agreement with the previously presented chromatographic results (Fig. 4A). Of note, the same approach applied to the bifonazole data did lead to different results. In this case, the decrease in the intensity of carotenoid spectra and increase in antifungal agent concentration were correlated solely with OD decrease (population size of the cells), and no change of the relative concentrations of the various carotenoids was observed. Therefore, the normalized spectra were identical over the entire bifonazole concentration range (Fig. S4), and no clustering was observed when PCA analysis was applied to these spectra (Fig. 5C).
To further investigate the influence of naftifine upon carotenoid synthesis, we performed experiments where the total carotenoids were monitored by absorbance at 512 nm after DMSO extraction from the cells and by OD 600nm to evaluate the cell density (as a marker of cell survival). Contrary to our expectation, hypoxia and UV exposure alone led to depigmentation and had only a weak influence upon cell survival, while the addition of naftifine further suppressed the carotenoid content. Repigmentation of cells treated solely with naftifine (Ctrl. + naftifine) occurred after reinoculation in naftifine-free media (Fig. 6A). A significant profile change in the UV-Vis spectra of the control, repigmented and hypoxia-treated cells was observed by PCA analysis (Fig. 5D). Strikingly, further chromatographic analysis revealed that the relative ratio of the determined carotenoids and ergosterol was very different among the control, naftifine-treated and repigmented forms. The torulene and C nd carotenoids were much lower in the repigmented forms than for the control batch, indicating an alteration of carotenogenesis after naftifine treatment, but β-carotene and ergosterol were much higher (Fig. 6B). This could be of particular importance for biotechnology involving carotenoid synthesis.
Following the chemical characterization, morphological and ultrastructural characteristics were evaluated using scanning and transmission electron microscopy. On the cell surface, a characteristic mucilage visible in scanning electron microscopy was observed (Fig. 7A). Ultrastructural characteristics of R. mucilaginosa cells include the electron-dense and lamellate cell wall, spherical to ovoid nucleus, plasmalemma, mitochondrion, endoplasmic reticulum, and lipid and glycogen accumulation in the cytoplasm. Numerous small lipid granules of uniform size were found in the cytoplasm of young cells, and larger lipid granules generated by fusion showed variable shapes in mature and senescent cells (Fig. 7B-D). Young R. mucilaginosa cells have a thin and electron-dense cell wall formed by elongation of the internal cell wall of the mother cell ( Fig. 7B and C). The depigmented cells of R. mucilaginosa had a low-electron-density cell wall without visible mucilage or lamellate structure and with large lipid granules (Fig. 7 E and F). R. mucilaginosa control colonies were mucoid and pigmented because the cells possess a carotenoid biosynthetic pathway 20,25 . Additionally, TEM revealed other ultrastructural characteristics as follows. The cells have lamellate and electron-dense cell walls and undergo enteroblastic budding similarly to basidiomycetous yeasts 36,37 . We suspected that the number of layers of cell wall was correlated with and proportional to the number of buds generated by a cell. In addition to these features, the cells of R. mucilaginosa accumulate different storage materials in the main phases of growth, such as glycogen and lipids, similarly to R. glutinis 38 .
Our results complete the list regarding the antifungal effects of naftifine, which provides good activity against Candida and Aspergillus species 39 Table 1. Chromatographic and spectral characteristics of the separated compounds. a The solvent strongly influences the spectrum profile; therefore, for reference, a solvent with similar polarity had to be used for a proper comparison. nd = not detected.  For the first time, we demonstrated that naftifine caused depigmentation of R. mucilaginosa colonies obtained on nutritive medium because the carotenoid biosynthetic pathway was affected. Our results are consistent with the results of other authors, who showed that naftifine has the ability to block the biosynthesis of carotenoids in Staphylococcus aureus by affecting the activity of the last enzyme in isoprenoid formation. Moreover, naftifine affected the budding of R. mucilaginosa cells, the synthesis of mucilage and the colour and lamellar structure of the cell wall. Bifonazole affected the biosynthetic pathway of ergosterol by inhibiting the demethylation of 4,4′,14-trimethylsterols or by inhibiting the microsomal HMG-CoA-reductase 42 . The carotenoid production was not affected in any way. Amphotericin B binds to ergosterol, determining the formation of pores in the cell membrane, and promotes the accumulation of ROS 43 . The MIC value determined against R. mucilaginosa was 1 µg/mL 44 .

Conclusions
Our research demonstrated that both naftifine (in Exoderil) and bifonazole (in Canespor) were effective antifungal agents for a R. mucilaginosa isolate causing onychomycosis, even though they acted differently. Of these fungicides, only naftifine affected the carotenoid synthesis pathway and produced depigmentation in R. mucilaginosa cells in a reversible fashion.
Furthermore, naftifine also reduced the relative carotenoid content after repigmentation. Future clinical studies are necessary to further investigate the implications of these findings.

Methods
Sample collection. Toenails affected by onychomycosis (Fig. 1A) were obtained from an 85-year-old female with chronic HBV hepatitis. After nail asepsis with 70% ethanol, the distal and lateral fragments of both toenail plates with subungual debris were removed with a sterile nail clipper.
The study was approved by the Ethics Committee of the Iuliu Hatieganu University of Medicine and Pharmacy, Cluj-Napoca (Romania) and an informed written consent was obtained from the patient before enrolling her in the study.
All experiments in this study were performed in accordance with relevant guidelines and regulations.
Fungal strain isolation and growth conditions. Small toenail fragments of 2-3 mm were disinfected in 20% ethanol for 1 minute and inoculated into SDA control media in Petri dishes 45 . Yeast growth was carried out using the standard method of triple culture (insemination of the inoculum in three points on the surface of SDA in Petri dishes) and incubated at 22 °C for 3 days. After incubation, the isolates were identified by morphological and cultural characteristics. All experiments were performed in triplicate. Fungal colonies obtained on SDA control media were also grown in Yeast Extract-Peptone-Dextrose (YPD) medium containing 0.5% peptone, 0.3% yeast extract, 0.5% glucose and 30 mg/L chloramphenicol. The media were sterilized at 121 °C for 30 minutes before chloramphenicol addition. Cultures of 10 mL were grown for three days at 22 °C in 50-mL Erlenmeyer flasks on a rotary shaker at 100 rpm (GFL Orbital Shaker 3017, Gesellschaft für Labortechnik mbH, Burgwedel, Germany). In all experiments, the relative yeast growth was evaluated by optical densities at 600 nm (OD 600nm ). Inoculation was performed by adding 50 µL of R. mucilaginosa inoculum at 2.4 OD 600nm (containing approximate 0.3 × 10 7 cells) in 10 mL culture media. Before running the experimental tests described in this study, preliminary studies involving OD 600nm and Raman profile monitoring were performed to create a growth chart for our experimental conditions, as well as measuring pH changes during the growth of the yeast until maturation. Antifungal product effects. Two pharmaceutical antifungal products, Exoderil (containing 10 mg/mL of naftifine as active compound, Sandoz GmbH Kundl Austria) and Canespor (containing 10 mg/mL of bifonazole as active compound, KVP Pharma + Veterinär Produkte GmbH, Kiel, Germany), were tested against R. mucilaginosa. Minimum inhibitory concentrations (MIC50), the minimum concentrations of compound required to inhibit 50% of the population in terms of OD 600nm values, for both antifungal agents was determined in liquid media and expressed in terms of the active compound concentration, followed by comparison with available data 41 . The tested concentrations were 0, 0.005, 0.01, 0.025, 0.05, 0.1, 0.2, 0.6, 1.2, 3.6, 10.8, 50, 100, 500, and 1000 mg/L for naftifine and 0, 0.015, 0.03, 0.06, 0.12, 1, 10, 50, 100, 500, and 1000 mg/L for bifonazole, respectively. Experiments were performed in duplicate.
The R. mucilaginosa strain isolated from the SDA control media was used as the inoculum for naftifine effects evaluation. The fungal depigmented colonies were obtained by using inoculum from control R. mucilaginosa isolate, exposed for 10 minutes to 1 mg/mL naftifine solution diluted with 20% ethanol, before application to nutritive medium, as inspired by practice recommendations of naftifine hydrochloride dosage and administration 47 . In another experimental variant, the R. mucilaginosa depigmented colonies, illustrated by photos, were obtained by including naftifine in small concentrations (0.1 mg/mL, 0.2 mg/mL) in SDA, before inoculation (Fig. 1C). The R. mucilaginosa repigmented colonies were obtained from fungal depigmented colonies by inoculation on SDA (Fig. 1D). All fungal colonies obtained on SDA in Petri dishes by method of triple cultures were incubated at 22 °C for 3 days, and all the experiments were performed in triplicate.
The concentration of 1.2 mg/L naftifine in YPD medium was chosen for subsequent investigations. An inoculum of 20 µL was used for 10 mL culture media. Two sets of experiments were carried out for R. mucilaginosa inoculated in YPD medium without naftifine and with naftifine. R. mucilaginosa previously treated with 1.2 mg/L naftifine was also used for inoculation in YPD medium without naftifine. For the first experiment, after inoculation the Erlenmeyer flasks were exposed to UV radiation for 40 minutes in a biological safety cabinet. For the second (hypoxic) experiment, flasks inoculated with R. mucilaginosa were sealed after argon was bubbled for 60 s in each flask. These tests were performed in quintuplicate. The results were reproduced in independent trials. Scientific REPORts | 7: 11125 | DOI:10.1038/s41598-017-11600-7 Carotenoid evaluation using UV-vis and resonant Raman spectroscopies. Cells from 2 mL of homogenous suspension were harvested and washed three times with 1 mL 0.9% NaCl solution; 100 µL of the final suspension in saline solution were used for Raman spectroscopy, while the other 900 µL were centrifuged and the saline solution was discarded. The pellet was resuspended in 1.8 mL of DMSO, immediately followed by vigorous shaking for complete cell disruption (transformation of the suspension into a clear solution) and carotenoid extraction. The clear solution was further centrifuged at 20000 g speed for cell debris deposition. The UV-vis spectra of the clear supernatant were measured between 200 and 800 nm using a Varian Cary 5000 UV-Vis-NIR Spectrophotometer (Agilent Technologies). Absorbance at 512 nm (maximum wavelength for these samples) was taken as a marker of the total carotenoid content.
Resonant Raman spectra were recorded using a confocal Renishaw inVia Reflex Raman Spectrometer at 5% of the 200 mW Cobolt Diode Pumped Solid State (DPSS) laser, emitting at 532 nm. The Raman back-scattered light was directed to a spectrometer equipped with 1800 lines/mm grating and a CCD detector. The spectral resolution was ~4 cm −1 . The instrument was calibrated prior to each experiment using a silicon sample as an internal standard. Cells in saline buffer were applied to poly-L-lysine adhesive slides (Polysine ™ Microscope Adhesion Slides, Erie Scientific via VWR) for efficient sample immobilization. The laser was focused on the sample using a 100x objective, and the spectra were recorded with 10-s exposure times and 2 accumulations. The experiments were done in triplicate.
Resonant Raman maps were obtained by raster-scanning the selected area with the 532-nm laser, focused on the sample using a 5x objective. The laser power was set to 10%. The exposure time for each spectrum was 2 s and the measurement step was 200 µm. The Raman signal contours map, as a function of the Raman intensity, is shown in Fig. 4C.
Chromatographic analysis. Both HPLC and HPTLC approaches were used for separating and evaluating the relative quantity of the carotenoids and ergosterol in R. mucilaginosa after DMSO cell disruption. An Agilent 1200 HPLC system (Waldbronn, Germany) equipped with an on-line vacuum degasser, quaternary pump, temperature-controlled sample tray, automatic injector, column thermostat compartment and DAD detector was employed, with a Nucleosil 100 C18 column (240 mm × 4.6 mm, 5 µm particle size) from Macherey-Nagel (Duren, Germany). The injection volume was 100 µL (0.2 µm filtered DMSO extract), the column temperature was set to 25 °C and the flow rate was 1.2 mL/minute. As a standard, β-carotene at a concentration of 1 mg/ mL was used. An isocratic elution was employed using acetonitrile:methanol:isopropanol at 85:10:5, as previously described 27 . UV-Vis detection of compounds was accomplished using the DAD detector to measure the entire spectrum in the 190-700 nm range every 1 s, and the chromatograms were monitored at 254, 450, 490 and 512 nm. The chromatograms were exported and the graphs were developed in Excel and Origin 6, where the integration of the peaks was also performed. HPTLC was performed on HP silica gel plates (Silica gel 60, Merck, Darmstadt, Germany) using hexane-acetone-methanol (10:15:75 v/v/v) as the mobile phase.
Electron microscopy. R. mucilaginosa cells were examined using scanning electron microscopy (SEM) with a JEOL JSM 5510 LV electron microscope and transmission electron microscopy (TEM) with a JEOL JEM 1010 electron microscope (Japan Electron Optics Laboratory Co., Tokyo, Japan). The samples were fixed in 2.7% glutaraldehyde (in phosphate-buffered saline for 90 minutes). For SEM, the samples were critical-point dried in liquid CO 2 , mounted on sticky carbon tabs and sputter-coated with gold (10 nm). For TEM, the fixed and dried samples were infiltrated with resin (Epon 812), deposited onto colloidal-carbon-coated copper grids and negatively stained with lead citrate and uranyl acetate 48 . The chemicals needed for electron microscopy experiments were purchased as follows: glutaraldehyde, resin (Epon 812), lead citrate, uranyl acetate, bismuth subnitrate (Electron Microscopy Sciences, Fort Washington, USA); sticky carbon tabs, colloidal carbon coated grids (Agar Scientific, Cambridge, England). Details regarding the sample work protocol for electron microscopy were presented in other reports 49 . Data analysis. Statistical analysis was performed using Statistica 7.0 for Windows (Stat-Soft, Inc., USA) and Excel. Principal component analysis (PCA) was applied to several data sets as indicated in the text using Statistica 7.0. T-tests were used to test the strength of association within the data. Where appropriate, using p < 0.05 as threshold for statistical significance, a statistical approach was formulated and the experimental data were also evaluated using one-way analysis of variance (ANOVA). The statistical parameters confirm the hypothesis that the differences between the results are either not significant (p > 0.05), significant (0.001 < p < 0.05) or highly significant (p < 0.001). The average of multiple measurements (duplicates, triplicates or more, depending on type of measurement) was used for plots, and the error bars are the standard error of the mean. For MIC determination, Origin 6 was used to fit the data and determine the MIC50. Data Availability. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.