Development of an Improved Carotenoid Extraction Method to Characterize the Carotenoid Composition under Oxidative Stress and Cold Temperature in the Rock Inhabiting Fungus Knufia petricola A95

Black yeasts are a highly specified group of fungi, which are characterized by a high resistance against stress factors. There are several factors enabling the cells to survive harsh environmental conditions. One aspect is the pigmentation, the melanin black yeasts often display a highly diverse carotenoid spectrum. Determination and characterization of carotenoids depend on an efficient extraction and separation, especially for black yeast, which is characterized by thick cell walls. Therefore, specific protocols are needed to ensure reliable analyses regarding stress responses in these fungi. Here we present both. First, we present a method to extract and analyze carotenoids and secondly we present the unusual carotenoid composition of the black yeast Knufia petricola A95. Mechanical treatment combined with an acetonitrile extraction gave us very good extraction rates with a high reproducibility. The presented extraction and elution protocol separates the main carotenoids (7) in K. petricola A95 and can be extended for the detection of additional carotenoids in other species. K. petricola A95 displays an unusual carotenoid composition, with mainly didehydrolycopene, torulene, and lycopene. The pigment composition varied in dependency to oxidative stress but remained relatively constant if the cells were cultivated under low temperature. Future experiments have to be carried out to determine if didehydrolycopene functions as a protective agent itself or if it serves as a precursor for antioxidative pigments like torulene and torularhodin, which could be produced after induction under stress conditions. Black yeasts are a promising source for carotenoid production and other substances. To unravel the potential of these fungi, new methods and studies are needed. The established protocol allows the determination of carotenoid composition in black yeasts.


Introduction
A wide variety of natural pigments are produced by a wide spectrum of organisms, including bacteria, plants, and fungi. Carotenoids are the main pigments besides the essential photosynthetic chlorophyll pigments. The carotenoid synthesis pathways of most organisms share two steps, which start with the synthesis of geranylgeranyl pyrophosphate (GGPP) from the head-to-tail condensation of four C 5 isoprene units and the tail-to-tail condensation of two GGPP units to produce the colorless difficult, since the absorption spectra depend on the solvent used [29] and the solvents can chemically react with pigments and such reactions can result in pigment derivatives [30].
Several methods were tested to increase extraction efficiency, mechanical or chemical treatment, and the use of specific solvents. The enhancement of pigment extraction was demonstrated by using mechanical or chemical treatment [31,32], ultra-sonication for cell disruption, and by using different extraction media [33]. Cell consumption is critical for quantitative pigment analysis. As [34,35] demonstrated the extraction efficiency depends on two parameters: an efficient cell disruption and a proper extraction medium. Mechanical treatment seems to be the preferred method for cells with thick cell walls [31,32]. Other parameters that influence pigment stability during extraction and therefore should be considered are pH-values, oxygen concentration, light and temperature.
The main fungal pigments are β-carotene, astaxanthin, lycopene, and neurosporene. Recently the production of torulene and torularhodin in fungi has gained some interest, especially as stress response to UV radiation [10,36,37]. The exact function of carotenoids in fungi is still under discussion, since mutants lacking carotenoids display the same growth rates as wild type strains. In the cases investigated, a lack of carotenoids has no apparent phenotypic consequences on growth or morphology in laboratory cultures. However, secondary effects of carotenoids, especially in regard to membrane integrity, are under discussion [6]. It is known in fungi that carotenoid formation is activated by blue light and H 2 O 2 , which generates oxidative stress [38]. The rock inhabiting fungus K. petricola A95 produces: β-carotene, γ-carotene, phytoene, torulene, and torularhodin and desiccation/rehydration stresses affect the formation of the colorless phytone as well as the other carotenoids [22]. A new pigment extraction protocol was used to investigate the carotenoid composition of the rock inhabiting yeast K. petricola A95 under control conditions but also under temperature and oxidative stress. This new, robust, and reproducible method allows quantitative analysis of carotenoids in black yeasts to be used to determine the adaptation responses to extreme environmental conditions in high stress resistant fungi.

Material and Methods
Cell cultures of K. petricola A95 were grown on MEA (malt extract agar) plates at 25 • C and 4 • C. Cell colonies were harvested after 7 days and freeze dried. Oxidative stress related pigment accumulation was tested with cell cultures grown (equal cell number) in liquid BG11 (cyanobacteria media,) media with 2% malt extract and 0.2% glucose and with subsequent addition of 20 mM H 2 O 2 for 2 h. K. petricola A95 was cultivated in BG11 media to investigate biofilm formation and interaction with cyanobacteria as described in [39], especially since such biofilms are exposed to harsh environmental conditions, leading to adaptations in pigment compositions in both organisms.
All cells were freeze dried after harvesting (Labconco FreeZone 2.5), dry weights of the cells were determined and the cells were homogenized. A mill for glass beads with integrated CO 2 cooling was used for the extraction of the pigments. The freeze dried material was transferred into a glass vessel with ground glass joint (Sartorius). The glass beads used had a mixing ratio of 1:3 (w/w), including 1 part of cells with a diameter of 1 mm and 3 parts of beads with an average diameter of 0.25 mm (Carl ROTH GmbH). Simultaneously the samples were disrupted with 100% HPLC grade acetonitrile (Sigma-Aldrich) for 3 min. To reduce the amount of fine particles in the supernatant, the samples were centrifuged at 20,800 g (Centrifuge 5417 C, Eppendorf) twice for 1 min after the disrupting procedure. The remaining pellets were tested for extraction efficiency by applying extraction solvents afterwards, to make sure, the pigment extraction was complete.
The determination of the different carotenoids was realized via HPLC (DIONEX, Ultimate 3000) A reversed phase C 18 column (300-5, C18-250 × 4 mm, from CS Chromatography Service) was used. A two eluent protocol was used with following gradient (Table 1)  Eluent A was composed of 80% acetonitrile (HPLC grade; Sigma-Aldrich) and 20% aqua dest. (v/v) and Eluent B was 100% ethyl acetate (HPLC grade; Sigma-Aldrich). The total protocol is 30 min long and was carried out with a flow rate of 0.8 mL/min. The used solvents were degassed for 15 min by ultrasonication prior use.
Pigment identification and quantification was established for lycopene, didehydrolycopene, dehydrolycopene, β-carotene, γ-carotene, torulene, and as well for torularhodin. Authentic standards were obtained for β-carotene and lycopene from Sigma-Aldrich. The quantification was based on the β-carotene calibration curve, only, because lycopene is not soluble in the extraction solvent. Lycopene was dissolved in 100% hexane (HPLC grade; Sigma-Aldrich). Calculation of pigment concentrations were based on β-carotene calibration and dry weight as reference values. Therefore, peak areas were related to β-carotene concentration and then divided by the specific extinction coefficients of the pigments in acetonitrile [40,41]. The pigments were detected at 440 nm for β-carotene, at 460 nm for of γ-carotene, lycopene and dehydrolycopene at 470 nm, torulene at 480 nm and torularhodin was detected at 497 nm. The dry weight was subsequently included. During the HPLC measurements, absorption spectra were ascertained by the HPLC-detector itself, supported by a tungsten lamp. The absorption spectra were determined from 402 to 762 nm. Furthermore, concentrations of purchased standards (β-carotene and lycopene) were verified by photometric measurements (Carl Zeiss Specord M500), to prevent overloading of the column. In all determined conditions three different biological replicates with a dry weight of about 0.1 g were analyzed. Pigment concentrations were subjected to the one-way analysis of variance (ANOVA) test in order to establish comparisons between different stress treatments. All datas generated or analyzed during this study are included in this published article.

Results
Pigment extraction was carried out with K. petricola A95 WT in comparison to a spontaneous melanin deficient mutant of K. petricola A95 (hereafter mdK) ( Figure 1). Eluent A was composed of 80% acetonitrile (HPLC grade; Sigma-Aldrich) and 20% aqua dest. (v/v) and Eluent B was 100% ethyl acetate (HPLC grade; Sigma-Aldrich). The total protocol is 30 min long and was carried out with a flow rate of 0.8 mL/min. The used solvents were degassed for 15 min by ultrasonication prior use.
Pigment identification and quantification was established for lycopene, didehydrolycopene, dehydrolycopene, β-carotene, γ-carotene, torulene, and as well for torularhodin. Authentic standards were obtained for β-carotene and lycopene from Sigma-Aldrich. The quantification was based on the β-carotene calibration curve, only, because lycopene is not soluble in the extraction solvent. Lycopene was dissolved in 100% hexane (HPLC grade; Sigma-Aldrich). Calculation of pigment concentrations were based on β-carotene calibration and dry weight as reference values. Therefore, peak areas were related to β-carotene concentration and then divided by the specific extinction coefficients of the pigments in acetonitrile [40,41]. The pigments were detected at 440 nm for β-carotene, at 460 nm for of γ-carotene, lycopene and dehydrolycopene at 470 nm, torulene at 480 nm and torularhodin was detected at 497 nm. The dry weight was subsequently included. During the HPLC measurements, absorption spectra were ascertained by the HPLC-detector itself, supported by a tungsten lamp. The absorption spectra were determined from 402 to 762 nm. Furthermore, concentrations of purchased standards (β-carotene and lycopene) were verified by photometric measurements (Carl Zeiss Specord M500), to prevent overloading of the column. In all determined conditions three different biological replicates with a dry weight of about 0.1 g were analyzed. Pigment concentrations were subjected to the one-way analysis of variance (ANOVA) test in order to establish comparisons between different stress treatments. All datas generated or analyzed during this study are included in this published article.

Results
Pigment extraction was carried out with K. petricola A95 WT in comparison to a spontaneous melanin deficient mutant of K. petricola A95 (hereafter mdK) ( Figure 1). The mutant allowed us to determine the quality of the carotenoid extraction since the containing carotenoids were visible in this mutant and not masked by melanin as in the WT of K. petricola A95. The mutant allowed us to determine the quality of the carotenoid extraction since the containing carotenoids were visible in this mutant and not masked by melanin as in the WT of K. petricola A95.
Pigment extraction was carried out for freeze dried cell pellets with a minimum weight of 15 mg. Torularhodin was the pigment with the lowest concentration and minimum weight of the used material was adjusted to determine the specific amount of torularhodin accordingly. The thick cell wall of K. petricola A95 caused the use of relative long cell disruption times, and the shorter extraction lead to a higher variation in extraction efficiency. Therefore to ensure quantitative pigment extraction, cell pellets were tested for the remaining pigments in a second extraction step with 100% acetonitrile and 100% hexane. No significant amounts of pigments were detectable via HPLC analysis (<2 µg/g dry weight), once the method was established. The only pigment left in the pellet was melanin in the WT. The efficient pigment extraction after cell disruption was confirmed by using mdK of K. petricola A95, leaving a pale pellet after extraction. Ideally, pigment extraction and cell disruption must be carried out under low temperatures and in the dark without oxygen to prevent pigment degradation.
To verify the usability of our system the described standards of all the pigments were tested in our system. We determined the retention times of β-carotene and lycopene. Although lycopene was not completely soluble in 100% acetonitrile, a mixture of 105:30:25 (v/v/v) hexane, acetone and acetonitrile was used for this standard. A second, isocratic 100% hexane, gradient was established and in some cases the isocratic separation with hexane was used for the re-extraction of the pellets.
Several extraction solvents as well as HPLC separation set-ups were tested. Chemical extraction, for 2 h at 60 • C with either dimethyl sulfoxide or n-propanol, was proven to be not successful. Although a high extraction efficiency for both solvents was expected according to [34] and [42] and DMSO was frequently used for studies with fungi [43]. In this study, acetone was used with the described mill with integrated CO 2 cooling combined with the Precellys 24 of Peqlab (at 6000 rpm for 5 up to 10 min) to ensure total cell disruption. In the latter cases the heating of the samples due to the rotation was the biggest problem and since sample cooling is crucial for stable pigment analysis, the mill was used exclusively. Furthermore, a solvent mixture of 40:20:40 (v/v/v) of acetonitrile, ethyl acetate, and n-propanol was tested, this mixture yielded good extraction results, but interacted negatively with the HPLC separation.
HPLC separation was tested with three eluents, acetonitrile/water, ethyl acetate, and ethyl acetate/acetone. Due to problems with a drifting baseline, eluents were reduced to eluent A (acetonitrile/distilled water) and eluent B (100% ethyl acetate). Eluent A was adapted, consisting of 80% acetonitrile (HPLC grade; Sigma-Aldrich)/20% distilled water. (v/v) was used. In combination with 100% acetonitrile as the extraction solvent the developed protocol ensures the detection of all pigments. All pigments show separate and distinct peaks. This was the most effective in extracting of the carotenoids from K. petricola A95 and was used as the established method. The total carotenoids in this approach were the highest, changing from, for example 153.7 µg g cdw −1 to 461.1 µg g cdw −1 .
In the final analysis, seven different pigments could be observed and quantified ( Figure 2) in K. petricola A95. The pigments are determined in the following order: dehyrolycopene (1), didehydrolycopene (2), which was always the largest peak, torulene (3), lycopene (4), γ-carotene (5) and β-carotene (6). The pigment torularhodin (7) was only detectable after prolonged incubation under low temperatures in trace amounts. Standards and absorption spectra were used to confirm the identity of the different pigments (Figures 2 and 3 and Table 2).   Short-term incubation (2 h) of WT K. petricola A95 under oxidative stress (treatment with 20 mM H2O2) showed an increase in concentration for several carotenoids compared to growth under normal conditions for the same time period (Figure 4). The concentrations of lycopene (38.7 to 113.2 µg g cdw −1 ), γ-carotene (6.6 to 18.2 µg g cdw −1 ) and dihydrolycopene (8.5 to 36.2 µg g cdw −1 ) increased strongly under the applied conditions (differences significant for p = 0.05). The concentrations of β-  (1), didehydrolycopene (2), torulene (3), lycopene (4), γ-carotene (5) and β-carotene (6) and torularhodin (7).  (4), γ-carotene (5) and β-carotene (6) and torularhodin (7).  Short-term incubation (2 h) of WT K. petricola A95 under oxidative stress (treatment with 20 mM H2O2) showed an increase in concentration for several carotenoids compared to growth under normal conditions for the same time period (Figure 4). The concentrations of lycopene (38.7 to 113.2 µg g cdw −1 ), γ-carotene (6.6 to 18.2 µg g cdw −1 ) and dihydrolycopene (8.5 to 36.2 µg g cdw −1 ) increased strongly under the applied conditions (differences significant for p = 0.05). The concentrations of β-  Short-term incubation (2 h) of WT K. petricola A95 under oxidative stress (treatment with 20 mM H 2 O 2 ) showed an increase in concentration for several carotenoids compared to growth under normal conditions for the same time period (Figure 4). The concentrations of lycopene (38.7 to 113.2 µg g cdw −1 ), γ-carotene (6.6 to 18.2 µg g cdw −1 ) and dihydrolycopene (8.5 to 36.2 µg g cdw −1 ) increased strongly under the applied conditions (differences significant for p = 0.05). The concentrations of β-carotene and torulene did not increase during the incubation and for didehydrolycopene an increase was detectable, however, the high variation between the samples did not allow us to draw conclusions. Didehydrolycopene showed in general the highest variation in pigment concentration in all samples.
carotene and torulene did not increase during the incubation and for didehydrolycopene an increase was detectable, however, the high variation between the samples did not allow us to draw conclusions. Didehydrolycopene showed in general the highest variation in pigment concentration in all samples. Incubation under low temperature did not influence the pigment composition ( Figure 5), compared to cultures grown under normal temperatures (the ANOVA Test showed no significant differences between the control samples and temperature stressed cells for all pigments). However, a strong increase in the didehydrolycopene concentration (up to 300 µg g cdw −1 ) under both conditions was visible after long term incubation. Therefore, the total measured pigment concentration increased to ~500 µg g cdw −1 . Identical results for all experiments are exhibited for the spontaneous mdK of K. petricola A95. The spontaneous mutant displays a similar amount of pigments per dry weight and the carotenoid concentrations did not change as a response to temperature stress (data not shown).  Incubation under low temperature did not influence the pigment composition ( Figure 5), compared to cultures grown under normal temperatures (the ANOVA Test showed no significant differences between the control samples and temperature stressed cells for all pigments). However, a strong increase in the didehydrolycopene concentration (up to 300 µg g cdw −1 ) under both conditions was visible after long term incubation. Therefore, the total measured pigment concentration increased to~500 µg g cdw −1 . Identical results for all experiments are exhibited for the spontaneous mdK of K. petricola A95. The spontaneous mutant displays a similar amount of pigments per dry weight and the carotenoid concentrations did not change as a response to temperature stress (data not shown).
J. Fungi 2018, 4, x FOR PEER REVIEW 7 of 12 carotene and torulene did not increase during the incubation and for didehydrolycopene an increase was detectable, however, the high variation between the samples did not allow us to draw conclusions. Didehydrolycopene showed in general the highest variation in pigment concentration in all samples. Incubation under low temperature did not influence the pigment composition ( Figure 5), compared to cultures grown under normal temperatures (the ANOVA Test showed no significant differences between the control samples and temperature stressed cells for all pigments). However, a strong increase in the didehydrolycopene concentration (up to 300 µg g cdw −1 ) under both conditions was visible after long term incubation. Therefore, the total measured pigment concentration increased to ~500 µg g cdw −1 . Identical results for all experiments are exhibited for the spontaneous mdK of K. petricola A95. The spontaneous mutant displays a similar amount of pigments per dry weight and the carotenoid concentrations did not change as a response to temperature stress (data not shown).

Discussion
For an efficient pigment analysis, two steps are important: First, a solvent for extraction should be selected that allows the quantitative determination of all relevant pigments. In specific cases and with unknown samples, a selection and/or a mixture of media should be used, like acetonitrile, ethyl acetate, isopropanol, and hexane. The most suitable solvents seemed to be isopropanol and hexane, which were used to verify the extraction efficiency (re-extraction of the pellet), but caused a baseline shift in the HPLC. It is noteworthy to state that DMSO is to our opinion not a suitable extraction solvent; the difficulties to remove DMSO from the system limit the application in an adequate eluent system. After centrifugation and removal, the supernatant residues of DMSO still remain in the sample, which is problematic for HPLC analyses. Additionally, the proposed DMSO based method of pigment extraction [37,44] should be tested accordingly, depending on the systems used, to ensure optimal pigment analysis results. Thus, DMSO extraction is always a time consuming method and the results are hardly comparable with other methods. Secondly an efficient cell rupture method under conditions that ensure the pigment stability should be selected. Black yeasts are characterized by a thick cell wall with high melanin concentration. In general pigment extraction should be performed under low oxygen, and if possible under oxygen free conditions. The major challenge in pigment analysis is to recover pigments with minimum risk of damage but with high efficiency. A mechanical treatment is needed in the most cases but the method applied should be tested carefully due to the high sensitivity of the molecules. Especially for black yeasts and other fungi with a thick cell wall, a mechanical cell disruption seems necessary and provides better results as a chemical treatment [31,32].
K. petricola A95 displayed an uncommon pigment composition; especially in the high concentrations of dihydrolycopene, didehydrolycopene and torulene, which are not so often described in fungi. A normally high abundance of carotene derivatives and lycopene can be observed in pigmented fungi [45,46]. Several yeasts showed a high torulene concentration but no dihydrolycopene was detected in yeast so far [46,47]. The abundance of didehydrolycopene and torulene indicates a similar production pathway as described for N. crassa, leading in this fungus to neurosporaxanthin [2,48], a carotenoid that was not detectable in K. petricola A95 under the applied conditions. It is noteworthy to mention that the determination of the abundance of astaxanthin and neurosporaxanthin is technically possible with the described protocol. The only pigment left in the pellet was melanin in the WT. Hence comparing the results with methods developed for pigments in algae or higher plants [26,27,31,[49][50][51] is not advisable due to the thick cultured cell wall of the supervised fungi. Although [49] compared different methods using maize seeds, they are not characterized by a thick cell wall like black yeast. Similar to our study, the decision regarding efficacy was made with respect to the color of the pellet. In general, these studies provide a wide method spectra and were useful regarding extraction, elution, and calibration of the authentic standards, see also [31,51]. Fungi pigments were mainly analyzed, focusing on certain valuable pigments only. Knowledge regarding pigment composition in fungi is needed for species description, however, such studies used standard methods and did not compare different methods [5][6][7]10,33,34,52]. Methodological overviews for fungi pigment extraction are still missing [38]. Therefore, the established method was compared to [19]. The combination of extraction media and the eluent system of [19] resulted in a strong injection peak in the presented HPLC system and was adjusted afterwards, concerning the different tested method combinations shown.
The detected carotenoids are consistent with [22], they found also carotene derivatives, torulene, and torularhodin. Our extraction method, especially the extraction medium and the mechanical treatment, gave us additionally the possibility to quantify lycopene and didehydrolycopene, unusual pigments in fungi and not often found in yeast strains.
To our opinion didehydrolycopene functions as a synthesis interface, comparable to β-carotene in photosynthetic organisms. Such a pool of didehydrolycopene can be converted fast to stress response pigments and helps cells to respond quickly to environmental stresses. The two main pigments in K. petricola A95, lycopene and didehydrolycopene, display much higher concentrations, as in other fungi detected [53] and β-carotene concentrations are lower compared to other fungi and yeasts. The detected lycopene derivatives are the major carotenoids in K. petricola A95. Future experiments must be carried out to find out if these pigments have a photoprotective role or if they function as a precursor for other pigments. In N. crassa the conversion of didehydrolycopene to torulene by cyclisation was described [54]. The already described torularhodin and torulene could be involved in photoprotection [6,13,37,55]. The antioxidant properties of torulene are attributed to its conjugated double bond system; in fact, torulene has more antioxidant efficiency than β-carotene, which presents less of a double bond on its chemical structure than torulene [56]. However, several other carotenoids were also described to have an antioxidant activity.
Another explanation for the stable carotenoid concentrations in K. petricola A95 would be that the mechanism of carotenoid action is more likely to consist of shielding sensitive molecules or organelles than of the neutralization of harmful oxidants [52]. Therefore, carotenoids do not play a major physiological role in fungal cells, but they may have beneficial effects under specific conditions. Several studies showed an increase in carotenoid concentration under stress [45,57]. The potential photoprotective pigment torulene [9,58] showed low concentrations in K. petricola A95, which would imply a less important role in K. petricola A95, but an inducible production under a stress situation out of the more stable dehydrolycopene seems possible. However, we could not detect an increase in torulene concentrations if the cells were exposed to oxidative or temperature stress. Such an induction was described for a black yeast under oxidative stress [59], in red yeast under salt stress [58], and a light dependent increase of the overall carotenoid concentration was described for N. crassa [38] incubated with H 2 O 2 . Such an inducible protective system would enable K. petricola A95 to react fast to stress situations. A similar system was described for the red yeast Dioszegia with the xanthophyll plectaniaxanthin [60]. The photoprotective carotenoid torularhodin was induced in K. petricola A95 under cold temperatures, but very low amounts and other stresses or longer incubation times have to be investigated to unravel the function of this pigment in K. petricola A95.
The major changes under the applied conditions were detected for lycopene and derivatives, which should therefore be considered as the most important pigments in K. petricola A95. The high concentration of the unusual carotenoid didehydrolycopene in K. petricola A95 should be investigated in detail, especially since the photoprotective role and the biotechnological potential of lycopene derivatives was described before [49]. Screening other black yeasts regarding their pigmentation could result in a wide diversity of pigments with multiple promising applications. Black fungi are remarkable in their stress resistance and we showed that their carotenoid pigmentation is complex and future studies have to be performed to determine the specific function of the specific pigments. The protocol presented will allow the quantitative analysis of black fungi, characterized by thick cell walls and melanin pigmentation, regarding stress response and adaptation to extreme environmental conditions. Future experiments have to unravel the function of melanin, carotenoids, and the cell wall structure regarding the stress responses in detail.

Conclusions
Extremotolerant and extremophile black yeasts are a promising source of pigments and other chemicals. New protocols and studies are needed to determine the capacity for production of these high stress resistance fungi. The established protocol allows the determination of carotenoid composition in black yeasts. Oxidative stress results in an adaptation in pigment composition. Future experiments have to be carried out to determine if didehydrolycopene functions as a protective agent itself or if it serves as a precursor for antioxidative pigments like torulene and torularhodin, which could be produced after induction under stress conditions. Author Contributions: J.T. and N.K. performed the growth experiments and drafted this manuscript. K.F. developed the HPLC methods and participated in the conception of the study, analysis of data and revision of the manuscript. J.T. and N.K. participated in the conception of the study, supervised the experiments, and revised the manuscript. All authors read and approved the final manuscript.
Funding: This work was supported financially by internal funds of the BAM.