Chemical characterization and biological effect of exopolysaccharides synthesized by Antarctic yeasts Cystobasidium ongulense AL101 and Leucosporidium yakuticum AL102 on murine innate immune cells

The current study aimed to investigate exopolysaccharides (EPSs) produced by two Antarctic yeasts isolated from soil and penguin feathers samples collected on Livingston Island (Antarctica). The strains were identified as belonging to the species Leucosporidium yakuticum (LY) and Cystobasidium ongulense (CO) based on molecular genetic analysis. The EPS production was investigated using submerged cultivation. Different chemical, chromatographic, and spectral analyses were employed to characterize EPSs. LY accumulated 5.5 g/L biomass and 4.0 g/L EPS after 120 h of cultivation, while CO synthesized 2.1 g/L EPS at the end of cultivation, and the biomass amount reached 5.5 g/L. LY-EPS was characterized by a higher total carbohydrate content (80%) and a lower protein content (18%) by comparison with CO-EPS (62%, 30%). The LY-EPS mainly consisted of mannose (90 mol%), whereas CO-EPS had also glucose, galactose, and small amounts of uronic acids (8–5 mol%). Spectral analyses (FT-IR and 1D, 2D NMR) revealed that LY-EPS comprised a typical β-(1 → 4)-mannan. Branched (hetero)mannan, together with β/α-glucans constituted the majority of CO-EPS. Unlike LY-EPS, which had a high percentage of high molecular weight populations, CO-EPS displayed a large quantity of lower molecular weight fractions and a higher degree of heterogeneity. LY-EPS (100 ng/mL) elevated significantly interferon gamma (IFN-γ) production in splenic murine macrophages and natural killer (NK) cells. The results indicated that newly identified EPSs might affect IFN-γ signaling and in turn, might enhance anti-infectious responses. The data obtained also revealed the potential of EPSs and yeasts for practical application in biochemical engineering and biotechnology.


Introduction
In recent years, scientists have concentrated their efforts on the exploration of the icy desert of Antarctica. This continent offers extreme weather events such as high UV, freezing temperatures, and strong winds to the microorganisms, hence they are forced into having an unusual metabolism that allows them to survive (Buzzini and Margesin 2014;Martorell et al. 2019). The extreme environmental conditions under which Antarctic yeasts actively grow and the lack of scientific research on their biosynthetic capabilities open up the possibility of isolating new molecules with bioactive properties such as polysaccharides. It is generally assumed that EPSs actively help yeasts to adapt to extreme environmental conditions. Some scientists have shared the opinion that EPSs protect the cell from biotic and abiotic stress (temperature, pH, or light intensity) (Donot et al. 2012;Ali et al. 2020). The exploration of the nature of Antarctic yeast-synthesized EPSs was the subject of previous studies by Pavlova et al. (2009Pavlova et al. ( , 2011 and Rusinova-Videva et al. (2011. These scientists have highlighted the fact that Antarctic yeasts infrequently synthesize pure homo-EPS. On the contrary, they produce hetero-ЕPSs comprised of mannose, commonly accompanied by glucose, galactose, arabinose, and non-carbohydrate constituents. For example, Cryptococcus laurentii AL 62 strain produce (xylo)mannan composed mainly of xylose and mannose, and a lower amount of glucose . Interestingly, an uncommon EPS is produced by a soil-inhabited strain Cryptococcus laurentii AL 100 . Its major sugar constituent is arabinose (61%) followed by mannose and glucose. Minor constituents are galactose and rhamnose. The structure and composition of EPSs seem to depend on the yeast strains and cultivation conditions used. Moreover, scientists have found evidence to suggest that polymers synthesized by Antarctic yeasts can be used as emulsifiers, stabilizers, and thickeners in the food and cosmetic industry (Pavlova 2014).
A large number of non-Antarctic yeasts produce EPSs (e.g. β-glucans, mannans) having prominent biological activities (antioxidant, anti-inflammatory, anti-tumor, immunomodulation). A Rhodotorula glutinins-produced EPS is a case in point. Its neutral sugar composition shows that mannose, glucose, and arabinose are the main sugar constituents. This EPS exhibits in vitro antioxidant activities, and what is more, it is capable of inhibiting the proliferation of the growth of colon carcinoma HCT-116 cells, and HAV virus (Ghada et al. 2012). Another representative of this genus, Rhodotorula mucilaginosa CICC 33,013 forms EPS (galactose, glucose, and mannose as main constituents) which exhibits radical scavenging and antitumor activities against human liver cancer cells (Ma et al. 2018). EPSs synthesized by Antarctic yeasts, on the other hand, do not affect the mouse monocyte-macrophage cell line viability and the proliferation of a human breast cancer cell line suggesting high compatibility and low levels of toxicity of these polymers (Poli et al. 2010). It seems that Antarctic yeasts are promising in having the capability to produce novel EPSs with new or improved properties allowing their application in different fields (nutritional, cosmetic, and pharmacological). This is why it is essential to screen more yeast strains that produce EPSs, and to study their physico-chemical and biological features, to find a better producer. Despite this, a large number of yeast genera have not been studied for their potential for biosynthesis of EPSs. Many structural characteristics of EPS constituents have not been investigated in detail. Moreover, very few scientists have devoted their attention to careful investigation of the biological effects of yeast-produced EPSs on innate immune cell activation (Reyes-Becerril et al. 2021). That is why, the main focus of our research was to investigate the potential of unexplored Antarctic yeasts to synthesize EPSs, the phylogenetical classification of the newly selected yeast strains, the characterization of the obtained biopolymers, and the identification of their biological effects on innate immune cells.

Antarctic yeasts
Two morphologically different yeast strains were isolated from soil and penguin feathers samples collected on the territory of the Bulgarian base on Livingston Island (Antarctica). The first strain, AL 101 was isolated from penguin feathers whereas the second (AL 102 ) was isolated from soil. The sample origin was selected because of the higher content of organic materials. These strains were stored in the Antarctic yeast collection of the Stephan Angeloff Institute of Microbiology, Bulgarian Academy of Sciences. Based on a genetic analysis of ITS1-5.8 S-ITS4 regions of rRNA (Valente et al. 1996;Rusinova-Videva et al. 2022) they were related to two genera -Cystobasidium and Leucosporidium. The phylogenetic analysis revealed that they belonged to Cystobasidium ongulense and Leucosporidium yakuticum respectively. The sequences were deposited in National Center for Biotechnology Information with the numbers SUB 11,505,954 and SUB 11,505,963, respectively.

Growth medium and cultivation conditions
The LB (Luria-Bertani) medium was used for the identification of the isolates. It was composed of 1.0% tryptone, 0.5% yeast extract, 1.0% glucose, 1.0% NaCl, 2.0% agar. This medium is used not only for bacterial cultivation but also for yeast cultivation (Panikov 2014). Before sterilizing the medium (20 min, 121 °C), its pH was adjusted to 7.5. The solid-phase cultivation was performed at 4 °C for 14-20 days to imitate the Antarctic temperature. It was experimentally proved that the best production of EPSs was at 22 °C for both strains, thus that temperature was used for further experiments. The nutrient medium used for submerged cultivation contained the following ingredients (w/v): 4% sucrose, 0.25% (NH 4 ) 2 SO 4 , 0.1% KH 2 PO 4 , 0.05% MgSO 4 × 7H 2 O, 0.01% NaCl, 0.01% CaCl 2 × 2H 2 O, 0.1% yeast extract (Sigma-Aldrich Chemie, Taufkirchen, Germany). The initial pH of the medium was adjusted to 5.3 and autoclaved for 30 min at 115 °C. The fermentation process was conducted in 500 mL flasks on a rotary shaker at 220 rpm, 22 °C for 120 h. The medium volume in the submerged fermentation was 10% of the volume of the flask. The inoculum was prepared using the same medium on a rotary shaker at 220 rpm, 22 °C for 48 h and it was added in an amount of 10% w/v (Rusinova-Videva et al. 2022).

Isolation of EPSs
Initially, the biomass was separated from the culture suspension by centrifugation at 5000×g for 25 min at 20 °C. The supernatant was then reduced threefold through a vacuum concentration at 50 °C (-0.1 MPa). To precipitate EPS, 96% (v/v) cold ethanol was added to the concentrated supernatant (1:3 v/v) and the mixture was incubated overnight at 4 °C. To recover the precipitate formed during storage, centrifugation at 5000×g was carried out for 20 min. The precipitate was then dissolved in ultrapure water and extensively dialyzed (VISKING®, SERVA Electrophoresis, Germany, MWCO 12-14 kDa) against distilled water for 72 h at 4 °C, with a periodic change of water. The dialyzed EPSs were centrifuged at room temperature, filtered through a Büchner funnel, and freeze-dried. The obtained precipitates were named CO-EPS (C. ongulense) and LY-EPS (L. yakuticum).

Determination of the biomass and EPS in the culture medium
The quantification of the biomass and EPS in the culture medium was gravimetrically carried out at the end of the cultivation (120 h). In brief, biomass was separated through centrifugation (5000 ×g, 30 min) from the culture suspension. The supernatant was used for the isolation of EPSs through precipitation with 96% ethanol (24 h, 4 °C). The precipitated EPS was dried at 105 °C to a constant weight. Biomass was washed twice with ultrapure water before drying at 105 °C until there was no change in its mass.

General analytical methods
A micro-Kjeldahl method was employed for the estimation of the crude protein content of the EPSs. The quantification of nitrogen expressed as an ammonia content of the digested sample was performed by acetylacetone-formaldehyde colorimetric method, using ammonium sulfate as a standard (NFSS 2010). The results were calculated using a nitrogento-protein conversion factor of 5.78. The total carbohydrate content of the EPSs was performed by the phenol-sulfuric acid method using mannose as a standard (DuBois et al. 1956). Initially, EPSs were pre-hydrolyzed in 72% (w/w) H 2 SO 4 (1 h, at 30 °C), and after dilution with water to 1 M H 2 SO 4 , hydrolysis was completed within the range of 3 h at 100 °C in a block heater (Stuart®, SBH200D). The obtained hydrolyzates were used as samples for analysis. The absorbance was measured at 490 nm. A part of the hydrolysate was taken for the analysis of the total uronic acid content. An automated 3-phenylphenol analysis was conducted by a continuous flow analyzer Skalar San ++ system (Skalar Analytical BV, Breda, the Netherlands), according to the instructions of the manufacturer. Absorption was measured at 530 nm and glucuronic acid (12.5-100.0 µg/mL) was used as a standard. The qualitative estimation of rare sugars was performed by the periodate thiobarbituric acid colorimetric method of Karkhanis et al. (1978) as described by Ognyanov et al. (2018). The acetyl content of EPSs was estimated photometrically by the hydroxamic acid reaction method of McComb & McCready (1957), using β-D-glucose pentaacetate (24-120 µg/mL) as a standard.

Monosaccharide composition analysis
Initially, 20 mg of EPS were hydrolyzed with 2 M trifluoroacetic acid (10 mL) for 1.5 h at 121 °C to release the monosaccharide constituents. Hydrolyzates were then vacuum-dried at 40 °C and re-dissolved in distilled water. This step was repeated twice to ensure the complete removal of trifluoroacetic acid. Finally, the hydrolyzates (10 mg/mL) were filtered (0.45 μm) and 10 µL were auto-injected into a Nexera-i LC2040C Plus UHPLC system (Shimadzu Corporation, Kyoto, Japan), coupled with a Zorbax Carbohydrate column (4.6 × 150 mm, 5 μm) and Zorbax Reliance Cartridge guard-column operating at 35 °C. The sample was eluted isocratically with a mobile phase composed of a mixture of acetonitrile/H 2 O (80:20 v/v) at a flow rate of 0.6 mL/min. The eluate was monitored using a refractive index detector RID-20 A (cell temperature 40 °C). The concentration of sugars in the sample was deduced using a calibration curve constructed by plotting the peak area (X-axis) against five different concentrations (Y-axis) for each sugar. The peak corresponding to different sugars in the sample was confirmed by a comparison of the retention time with that of the standards. D(+) Man (99%), L(-)-Fuc (99%), and GlcA (≥ 98+%) were purchased from Alfa Aesar, while D(+) Xyl (≥ 99%), D(+) Gal (≥ 99%), D(+) Glc (≥ 99.5%) were bought from Sigma-Aldrich.

Molecular weight distribution analysis
Initially, the EPSs (2 mg/mL) were completely dissolved in distilled water for 24 h before beginning molecular weight distribution analysis. The analysis of EPSs was carried out on a Nexera-i LC2040C Plus UHPLC system (Shimadzu Corporation, Kyoto, Japan), coupled with a RID-20 A detector, using a Bio SEC-3 column (4.6 × 300 mm, 300Å, 3 μm, Agilent). Ten microlitres of the filtrated (0.45 μm) solution of the sample were auto-injected and eluted isocratically at 30 °C with a mobile phase composed of 150 mM NaH 2 PO 4 (pH 7.0), employing a flow rate of 0.5 mL/min. Pullulan standards (Shodex standard P-82 kit, Showa Denko, Japan) with molecular weights in the range of 0.59 × 10 4 to 78.8 × 10 4 g/mol were used for the construction of a logarithm standard curve.

Fourier transform infrared (FT-IR) spectroscopy
The FT-IR spectra of the EPS samples (2 mg) were recorded in the region 4000-500 cm −1 using the attenuated total reflection technique on Tenzor 27 (Bruker, Germany), controlled by OPUS 8.7. software. The two spectra were analyzed in Spectragryph software (Dr. Friedrich Menges).

Nuclear magnetic resonance (NMR) spectroscopy
For better exchange of H 2 O with D 2 O required by NMR analysis, samples (30 mg) were dissolved in 700 µL D 2 O and lyophilized. They were then re-dissolved in 700 µL D 2 O and transferred with Pasteur pipettes into NMR cuvettes. 1 H, 13 C, DEPT-135 and 13 C/ 1 H HSQC, 1 H/ 1 H COSY, DIPSI, ROESY, and 13 C/ 1 H HMBC spectra were obtained on a Bruker Avance II + 600 spectrometer (Bruker, Germany), equipped with Bruker TopSpin ™ software operating at a temperature of 60 °C. Sodium 4,4-dimethyl-4-silapentanesulfonate (DSS) was used as an internal standard.
Splenocytes or BM cells were added at a concentration of 1 × 10 6 /mL in a volume of 250 µL into a 24-well plate (TPP, Switzerland). The CO-EPS and LY-EPS preparations, and positive control of lipopolysaccharides (LPS) (from Escherichia coli O55:B5; Sigma Aldrich), and zymosan (Z4250, Zymosan A from Saccharomyces cerevisiae; Sigma Aldrich) were added at volume 250 µL/well at increasing concentrations varied from 100 ng/mL to 100 µg/mL. The concentration range was the final concentration in the wells. Cells were cultured for 24 h at 37 °C, with 5% CO 2 (Heraeus, Germany). At the last 2 h of culture, 2 µM monensin sodium salt (Sigma Aldrich) was added at the volume of 100 µL/well to block cytokine secretion.

Flow cytometry analysis
After cell cultivations in the presence/absence of the EPS preparations, cells were collected, washed 2 times with PBS, counted, and resuspended in 1 × 10 5 /mL in FACS buffer (PBS, 2% bovine serum albumin, 1 mM EDTA, 0.1% sodium azide, all from Sigma-Aldrich, Germany). Cells were incubated for 20 min, in a dark place, 4 °C with 0.1 µg/ mL of the following antibodies: anti-CD335 labelled with FITC, anti-TRAIL labelled with PE, anti-F4/80 labelled with PE, anti-TLR24 labelled with PE. Cells were washed 2 times with PBS and then fixed with 4% paraformaldehyde/ PBS (Biolegend, UK) for 10 min at room temperature. For intracellular flow cytometry, after 2 times washing in PBS, the cells were treated with 500 µL of permeabilization buffer (Biolegend, UK) and incubated for 40 min, in a dark place, at 4 °C with 0.1 µg/mL. Anti-IFN-γ antibody labelled with APC or Cy7 (Biolegend, UK). Cells were washed 4 times with PBS and subjected to analysis at a BSR II flow cytometer (Beckton Dickinson, USA) using DIVA 6.0 software (Beckton Dickinson, USA).

Statistics
The flow cytometry data are expressed as mean ± standard error. The significant differences between the groups were evaluated by a one-way analysis of variance (ANOVA) test. A p-value less than 0.05 was considered significant.
All yeast cultivations were performed in duplicates. The HPLC analyses were performed at least in duplicates whereas the other analyses were run at least in triplicates. Results were expressed as mean values ± standard deviations. ANOVA and Student's t-test were used to evaluate the differences in the mean between groups. P values less than 0.05 were considered to be significant. Microsoft Excel, 2016 (Microsoft Corporation, Redmond, USA) was used in the analyses.

Identification and cultivation of yeast strains
Two different strains were identified as Cystobasidium ongulense (CO) and Leucosporidium yakuticum (LY). They differed in respect of size and morphological characteristics. CO colonies were orange in colour and the cells were elliptical in size of 4.8 × 2.9 μm, while the colonies of LY were characterized in white colour and the cells had a smaller size (3.9 × 2.6 μm) (Fig. S1 a, b). Figure 1 represents the yeast growth, accumulation of EPSs, and the change of pH of the culture medium during the fermentation process. CO followed the standard course of exponential growth, although it reached slowly the stationary phase after 72 h (Fig. 1a). By contrast, LY reached the stationary phase between 48 and 72 h (Fig. 1b). The amount of accumulated biomass reached the highest level of 5.4 g/L (CO) and 5.65 g/L (LY) at 120th h of cultivation. As can be seen from these figures, the accumulation of biomass was paralleled by an increase in the quantity of EPS till the end of 24 h, but afterward, the accumulation pattern of EPS changed significantly for both species. Figure 1a shows that the largest quantity of CO-synthesized EPS (2.1 g/L) was determined at 72 h. A significant proportion of CO-EPS (> 77%) was actively synthesized during the first 24 h of cultivation. Until the end of cultivation, its amount decreased by 38%, but this did not change the course of yeast growth. It can be seen that LY accumulated a higher amount of EPS (4.0-4.5 g/L) at a later stage of cultivation (96-120th h), although a higher proportion of EPS was formed up to 24th h (> 66%) (Fig. 1b). Interestingly, after a short period of stagnation (24-55 h), the content of EPS increased smoothly till the end of the cultivation process. However, this period was accompanied by excessive yeast growth. Concerning EPSs, CO-EPS yielded the lowest amount of EPS (2.1 g/L at 48 h) in comparison with LY-EPS (4.5 g/L), although it accumulated a higher amount of biomass.
After starting the cultivation, pH began to decrease from 5.3 to 2.0 for 48 h. A sharp decrease in pH was observed in the first 24 h for rapidly growing LY, while the CO's development caused the pH value to fall slowly to its lowest level for 48 h.

Chemical characterization of EPSs
The yield of EPSs is shown in Table 1, where monosaccharide composition and other chemical characteristics are included. LY seemed to be a better producer of EPS in comparison with CO strain. Interestingly, three times as much EPS was formed by LY as CO for the same period of cultivation (120 h). The total carbohydrate content of LY-formed EPS represented a larger proportion of the dry matter content (80%), while the other producer accumulated polymer characterized by a lower amount of carbohydrates (62%). An interesting finding of the study was that protein constituted an appreciable amount of CO-formed polymer (30%). By contrast, LY-EPS was accompanied by a smaller amount of protein (18%). It seems that the investigated EPSs were of neutral type because neutral sugars accounted for 66% (w/w) and more of the total sugar content, while acidic components did not exceed 6% (w/w). Next to mannose (30% w/w), CO-EPS was also composed of glucose, galactose, and minor quantities of uronic acids suggesting that this polymer was of hetero-mannan type. Compared to CO-EPS, LY-EPS was found to be highly enriched in mannose which may be accounted for the presence of a mannan-type PS. This sugar made up nearly 95% of total neutral sugars, whereas the uronide-containing components were found in a fairly low percentage. As can be seen, EPSs tested negative for rare sugars, because a less typical orange colour was observed instead of pink. It is interesting to note that some sugar residues that constituted EPS were O-acetylated to a different extent. Acetyl groups were found in a low amount in LY-EPS, but they were present in a much higher amount in CO-EPS.

Molecular weight distribution of EPSs
To estimate the molecular weight distribution of EPSs, an HPSEC analysis was employed and the resulting elution profiles are shown in Fig. 2. High molecular weight fragments in LY-EPS were recognized. They were eluted in the range of RT 5.0 and 7.0 min (Fig. 2a). The main peak covered the mass range between > 47 × 10 4 g/mol and 70 × 10 4 g/mol. The high molecular weight fractions (RT 5.5-7.0 min) occupied a higher percentage (64%) of the total (100%) peak area, hence the large percentage of EPS. As shown in Fig. 2a, a small amount of EPS fraction (26%) was eluted at an RT of 7.5-10 min, covering a mass range between 10 × 10 4 g/mol and 1 × 10 4 g/mol. There was a marked difference in the CO-EPS elution pattern. CO-EPS consisted of two distinct fractions with different distributions of molecular weights and higher heterogeneity. The first main fraction was eluted early between 5.5 and 7.5 min. It comprised about 28% of the total peak area (100%), and therefore a smaller percentage of EPS. It covered the mass range of 78 × 10 4 and 21 × 10 4 g/ mol. Next to those fractions, a second main peak was eluted (RT 8.5-11.0 min), which represented fractions composed of lower molecular weight fragments. Different peaks covered the range of 11.2 × 10 4 and 1.2 × 10 4 g/mol occupying 3%, 8%, and 61% of the total peak area suggesting that a large quantity of lower molecular weight fractions was present.

FT-IR spectroscopy of EPSs
The FT-IR spectrum of LY-EPS and CO-EPS are shown in Fig. 3. It was very clear that the spectra were typical for polysaccharides because a band at about 3300 cm −1 and two bands at about 2900 cm −1 , assigned to O-H, C-H, and C-H 2 stretching respectively, were observed. However, the band at 3340 cm −1 due to O-H stretching could not be distinguished from that showing N-H stretching (hydrogen-bonded N-H 2 group). Several bands can be assigned to the amide group vibrations bearing in mind a higher protein content present in the EPSs. The bands at 1542 and 1638 cm −1 were assigned for δ(N-H) and ν(C-N) of amide II and N-linked C=O of amide I structure vibrations (Hamidi et al. 2020). The absorption bands at 1638 and 1411 cm −1 could be also attributed to δ s (CH 2 ) and δ s (CH 3 ) of proteins rather than the ionized carboxyl groups of uronic acids because they were found in smaller amounts (Synytsya et al. 2003). The C=O stretching vibration of the O-acetyl group (ester) and τ(CH 2 ) gave rise to absorption at 1246 cm −1 together with a weak band at 1720 cm −1 . Moreover, bands at 1246 and 1312 cm −1 may be caused by amide III vibrations (δ(N-H) and ν(C-N)) of protein components. The band at 1375 cm −1 , in general, assigned to the C-H bending vibration of CH 3 group and ω(CH 2 ), may originate from C-H vibrations in monosaccharide residues. The presence of bands at 1151 cm −1 and 1061 cm −1 corresponded to O-C-O, C-O, and C-C stretching of glycosidic bond vibration of mannose-containing polysaccharides. Other bands encountered with mannan-type PS were: 1024, 943, and very specific signals at 872 and near 807 cm −1 which could be assigned to C-O, C-C stretching, C1-H bending, and ring vibration (C-O, C-C) of monosaccharide rings respectively. Further, the IR-FT spectrum showed the presence of bands for β-anomer configurations  (Kato et al. 1973). Based on chemical composition and FT-IR spectral analysis, it can be supposed that LY-EPS comprised a typical β-mannan PS. CO-EPS, on the other hand, seemed to be hetero-mannan and β-glucan PSs.
Interestingly Additionally, there was a correlation between the proton and carbon from the CH 3 group of acetyl esters at 2.16-2.20/21.1 ppm in the HSQC spectrum (Fig. 4). It was speculated that acetyl groups were localized at O-2 and/or O-3 of Man residues, as it is found in different β-glucomannans (Hannuksela and du Penhoat 2004;Makarova et al. 2018). Apart from that, the minor signals at 4.59 and 4.51 ppm were attributed to H-1 of the terminal Xyl residue, linked to β-D-Manp (Makarova et al. 2013;Shakhmatov & Makarova 2022).

Biological effect of exopolysaccharides on innate immune cells
Furthermore, we focused our attention on investigating the biological effect of EPS on innate immune cells. Two cell cultures were used as a source of innate immune cells: bone marrow and spleen. In BM we identified monocytes with the origin of granulocyte precursors as Ly6C + cells. In the spleen, we followed the effect of EPS on macrophages positive for F4/80 and on NK cells positive for CD335. The intracellular level of TNF-α in BM monocytes stimulated in the presence of the EPS preparations and two strong activators zymosan (TLR2 agonist) and LPS (TLR4 agonist) was determined. Unstimulated monocytes (control group) showed the highest percentage of TNF-α positive cells in comparison to all groups treated with preparations, LPS, or zymosan (Fig. 6a), which was indicative of significant TNF-α storage. The stimulation with zymosan, TLR2 agonist, reduced the intracellular amount of the cytokine in a dose-dependent manner, and thus the TNF-α positive cells decreased. We suggested that the cytokine can be secreted by the intracellular stores during the stimulation period of 24 h (a period when the cytokine secretion was unblocked by monensin), thus reducing the number of cells positive for the cytokine. The effect of LPS was stronger than that of zymosan showing the depletion of TNF-α positive cells. We observed that all EPS solutions significantly reduced TNFα positive monocytes even at concentrations of 100 ng/ml while LPS had a strong effect on TNF-α release.
F4/80 is a marker for tissue-resident macrophages. Tissue macrophages have an important role in the regulation of granulopoieses in BM and emergency granulopoiesis in organs during acute inflammation or tumor growth. We observed that the percentage of TNF-α positive splenic macrophages was higher in the control group and the group treated with 100 µg/mL LPS (Fig. 6b). It is possible that in macrophages a high concentration of LPS-induced tolerogenic signals and a smaller amount of TNF-α was released. A large number of TNF-α positive cells were found in the LY-EPS-treated group (100 ng/ml) compared with the CO-EPS-treated groups. Our data showed that either LY-EPS stopped the TNF-α releasing by the macrophages or it induced de novo synthesis of the cytokine that was high enough to fulfill the depleted intracellular store of the cytokine. Moreover, we also evaluated the effect of EPS preparations on IFN-γ production in macrophages (Fig. 6c). It was observed that LY-EPS increased IFN-γ production in a dosedependent manner. At the lowest concentration of 100 ng/ mL LY-EPS elevated significantly the percentage of IFN-γ producing cells. It has been shown that NK cells might be sensitive to EPS and that LPS can induce IFN-γ production in human NK cells (Kanevskiy et al. 2013). Therefore, we isolated splenocytes and stimulated them for 24 h with EPS and LPS. Our results verified that LPS induced IFN-γ production in murine NK cells. We found that a low concentration of LY-EPS increased significantly IFN-γ production by comparison with LPS. However, the CO-EPS sample failed to affect IFN-γ production (Fig. 6d).
It may be that EPSs affect TLRs expression. TLR2 is an innate receptor that interacts with a diverse range of microbial pathogen-associated molecular patterns (PAMPs) from bacteria (Gram-positive and Gram-negative), fungi, parasites, and viruses. These PAMPs include cell-wall components, such as lipoproteins, lipoteichoic acid (LTA; Grampositive bacteria only), lipoarabinomannan (mycobacteria only), and zymosan (yeasts). TLR2 can be also a ligand for Fig. 6 Effect of exopolysaccharide solutions (L. yakuticum-EPS and C. ongulense -EPS) on intracellular a TNF-α production in Ly6C + monocytes; b TNF-α production in F4/80 macrophages; c IFN-g production in F4/80 macrophages; d IFN-g production by splenic NK cells; e on TLR2 expression by Ly6C + monocytes; f on TRAIL expression by splenic NK cells. The data represent mean ± SE, *p < 0.05, **p < 0.01, ***p < 0.001 when compare to control; # p < 0.05, ## p < 0.01, ### p < 0.001 when compare to group with stimuli LPS or Zy, Anova test EPS from Lactobacillus rhamnosus GG that can alleviate adipocyte functions (Zhang et al. 2016). EPS from Lactobacillus plantarum can induce apoptosis via the engagement of TLR2 and further induction of FAS/FASL apoptosis (Zhou et al. 2017). Thus, we studied the level of TLR2 in monocytes after treatment with EPS solutions. It was found that zymosan induced expression of TLR2 at a low concentration, whereas, at a high dose, it decreased TLR2, because it induces internalization of the receptor. We observed that none of the EPS solutions was able to induce TLR2 expression similarly to LPS (Fig. 6e). The transmembrane protein TRAIL is known to induce apoptosis in tumor cells (Wiley et al. 1995). The results of our experiment showed that its expression level after treatment with LY-EPS (10 µg/mL) and CO-EPS (100 ng/mL and 100 µg/mL) were higher than those in the control (Fig. 6f).

Discussion
As stated in the Introduction, our main aim was to investigate two yeast strains collected in Livingston Island (Antarctica) and to provide a useful piece of information about the production and chemical characteristics of EPS.
According to our investigation, yeast producers were identified as CO and LY. For the first time, CO is described as a soil inhabitant isolated from East Ongul Island (East Antarctica) in 2017. Other species of this genus have also been found in the high mountainous parts of China and Italy (Tsuji et al. 2017). CO has been reported as a sucroseconsuming strain that can produce valuable enzymes (esterase, lipase, and β-glucosidase) (Tsuji et al. 2017). LY, also known as Rhodotorula yakutica and Leucosporidiella yakutica, is described for the first time by Li et al. (2020). Species of Leucosporidium's genus have been found in permafrost in Antarctica (Vishniac 2006). L. antarcticum is known to synthesize proteinase and cold-active glucosidase and β-fructofuranosidase (Turkiewicz et al. 2003(Turkiewicz et al. , 2005. However, except for earlier findings about EPS-synthesizing L. scottii (Rusinova-Videva et al. 2019), we did not find any other reports on EPS from another genus member. Both yeast producers have not been studied for their ability to synthesize EPS.
In general, it seems that the yields of EPS and biomass are considerably influenced by differences in the metabolism of yeasts, carbon source, and cultivation conditions. By comparison, our results for the yield of biomass of LY and CO (4-5 g/L) are dissimilar to Rusinova-Videva et al. (2020) who reported a higher yield of biomass (7.0 g/L). However, a key difference is the use of another yeast strain (Cryptococcus laurentii AL 65 ). Another example of this is yeast strain (C. laurentii AL 100 ) cultivated on a sucrosecontaining medium which also accumulated not more than 3 g/L biomass. Concerning EPS, there is not much difference between the yield of LY-EPS and that of EPS produced by C. laurentii AL 65 (3 g/L at 48 h) strain cultivated on a sucrose-containing (2%) culture medium . Interestingly, LY-EPS is comparable in yield of EPSs formed by other soil-inhabited producers such as Vishniacozyma victoriae and Tremellomycetes sp. (4.0-4.5 g/L), placed under the same conditions (Rusinova-Videva et al. 2022). However, there was a significant difference in terms of yield between LY-EPS and CO-EPS. Bearing in mind that CO differed from LY in respect of the habitat (penguin feathers and soil), we should not underestimate the importance of strain origin as regards EPS yield.
Further, our current findings about the decrease in pH during the first hours of cultivation are in complete agreement with previous studies (Pavlova et al. 2004). It may be that lower pH gives yeast strains a competitive advantage over other microorganisms.
Broadly speaking, the chemical composition and molecular weight distribution of yeast EPSs are considered essential for the explanation of the biological activity and functional properties of PSs. In the current study, we succeeded in revealing that both EPSs were rich in mannose and protein. Moreover, spectral analyses revealed that LY-EPS was formed mostly of β-(1→4)-mannans accompanied by protein (e.g., mannoprotein), whereas CO secreted into the culture medium EPS comprised mainly of a mixture of hetero-mannans, β-glucan, and protein (e.g. a galactoglucomannan protein fraction). It is a well-documented fact that carbohydrate polymers such as glucans, mannans, and chitin, together with lipids and proteins are major constituents of yeast cell walls . However, there is a wide variety of EPSs characterized by different compositions depending on the yeast strain and the composition of the culture medium. For example, Han et al. (2018) demonstrated that Sporidiobolus pararoseus JD-2-excreted EPS was composed of galactose, glucose, and mannose having a ratio between them of 16:8:1. Uronic acids were not present in the EPS. In our recent study, on the other hand, Vishniacozyma victoriae and Tremellomycetes sp.-derived EPSs were composed mainly of mannose (37 mol%), while galactose and uronic acids occurred in smaller amounts (Rusinova-Videva et al. 2022). Moreover, an earlier study showed that Sporobolomyces salmonicolor AL 1 strain produced EPS consisted not only of mannose (43%) and glucose (54%) but also of a small amount of fucose (3.3%). It is interesting to note that EPS composition could be changed when yeast cells are cultivated under stress conditions. For example, the basic sugar constituents of EPS produced by Cryptococcus laurentii under salt-stress conditions were mannose, xylose, and galactose in the proportion of 10:6.1:1.1, while the EPS formed under optimal conditions was characterized by a molar ratio of 10:3.4:1.3 (Breierová et al. 2005). It is worth mentioning that CO-EPS and to a greater extent LY-EPS contained acetyl groups. This is in line with previous findings about an acetylated hetero-mannan (3-10%) (Van Bogaert et al. 2009).
A recent review of the scientific literature on this topic highlights that there is a great variety not only of constituents of EPS but also of glycosidic linkage patterns. For example, some fungal and yeast α-1,6-mannans are branched at O-2 with several α-D-Manp residues, which could be additionally substituted at O-3 with terminal α-D-Manp units (Gómez-Miranda et al. 2004;Galinari et al. 2017). Interestingly, Matsuo et al. (2000) isolated acetylated 1,3,4-β-Dhomomannan from Leptospira biflexa and they verified by a series of experiments Man O-3 substitution. As reported above β-1,4-mannans (constituted LY-EPS) could be also substituted with Fuc units. This lends support to previous findings of a study by Yamada et al. (1982) who isolated an acetylated PS from the fungus Absidia cylindrospora. The PS was composed of α-1,6-mannan branched at O-3 with α-L-Fucp.
The unique properties of microbial PSs determine the possibility of their use for nutritional, cosmetic, and pharmacological applications in the role of immunostimulants, antitumor, and antioxidant drugs (Freitas et al. 2011;Ghada et al. 2012;Ma et al. 2018;Ragavan and Das 2019;Seveiri et al. 2019;Hao et al. 2020;Hristova et al. 2021). Moreover, polysaccharides' diversity in terms of species, chain length, molecular weight, and conformation of monosaccharide residues is crucial for the diversity in their biological role (Yang and Zhang 2009;Zhou et al. 2019). Some β-glucans and their conjugates, for example, are known as biological response modifiers (Leung et al. 2006) and effective immunostimulant and antigenic carriers (Sanchez et al. 2021). They bind to innate immune receptors triggering immune cell activation and further development of immune responses (Murphy et al. 2007). In addition, some β-glucans also exhibited antioxidant capacity, as was reported by Reyes-Becerril et al. (2021).
There was limited knowledge of the biological activity of EPSs synthesized by Antarctic yeast in the literature. For this reason, our research focused on the effect of EPS on innate immune cells. We found evidence to support the view that either EPS affected other TLRs (TLR2 expression by Ly6C + monocytes) or induced internalization of the receptor affecting its turnover on the cell surface. We also observed that LY-EPS and CO-EPS sustained a low percentage of TNF-α positive BM monocytes even at concentrations of 100 ng/mL. However, only LY-EPS was able to increase TNF-α intracellular levels in tissue-resident splenic macrophages. It was very likely that other preparations might induce tonic signals that failed to elevate TNF-α production. Furthermore, the significant influence of IFN-γ on F4/80 macrophages allowed us to speculate that LY-EPS affected IFN-γ signaling, probably via engagement with innate receptors and pathways leading to activation of IFN-γ production. This is consistent with previous findings in the literature. For example, Lackovic et al. (1970) reported that some mannans and mannan-protein complexes stimulate the release of interferon. The authors had reason to believe that mannans stimulated the release of IFN-γ by a mechanism similar to that of endotoxins but exhibiting minimal toxicity. Moreover, our findings on the immune activity and structure of the EPSs would seem to suggest that the LY-EPS having a β-mannan-type structure exhibited a stronger effect on IFN-γ in specific immune cells in contrast to the more heterogeneous CO-EPS.
Another very interesting question to discuss is the relationship between the structural features of EPSs and the cold adaptation of yeasts. It seems that EPSs, together with antifreeze proteins, play a leading role in the adaptation of yeast species to extreme environmental conditions. In addition to cell adhesion, they favor the retention of liquid water and the concentration of nutrients. They probably act as non-penetrating cryoprotectants like antifreeze proteins and exo-glycoproteins Donot et al. 2012). Several authors have attempted to reveal the dominant motif responsible for the antifreeze activities of proteins -one of the cold adaptation mechanisms. So far, however, little has been achieved. For example, Kim et al. (2014) reported that an antifreeze protein isolated from a cold-adapted psychrophilic yeast Glaciozyma sp. AY30 was composed of 261 amino acids with an N-terminal signal sequence and one N-glycosylation site (Asn185). Interestingly enough, the authors have noted that the different types of antifreeze proteins had different primary, secondary, and tertiary structures, but their activity did not change significantly. As for EPSs, the question remains unanswered, and what is more, with a lack of empirical shreds of evidence, much speculation is running. There is no doubt that mannoproteins are the basic constituent of EPSs produced not only by most cold-adapted yeasts (30-90 mol%) but also by a large number of enological Saccharomyces cerevisiae strains (Bicca et al. 2022). As was mentioned above in the Discussion, a great variety of glycosyl linkage patterns for mannose and other sugars (if present) in the purified EPSs exist. Nevertheless, we can only speculate that for a better cold adaptation of yeasts the glycosidic linkages of the mannose units of the synthesized EPSs should predominantly be of the β-type. This statement might not be generalized to all cases because many EPSs excreted by cold-adapted yeasts have also an α-type. Moreover, it should be mentioned that the conformation of the structural unit and glycosidic linkage type in the chain determines the chain conformation of a polysaccharide (Belitz et al. 2009). For example, ribbon-type conformation is typical for β-1,4-linked glycans, which are usually insoluble in water. This can exert a strong influence on the physicochemical (functional) properties and antifreeze activity of EPS, and thus on the mechanism of adaptation. In our opinion, due to the great diversity of yeast strains, cultivation conditions, EPSs composition, and unknown mechanisms of cold adaptation, scientists still do not fully understand which structural motif of EPS (if any) plays a leading role in contributing to cold adaptation.
In conclusion, we have succeeded in providing valuable insight into the composition of EPSs synthesized by poorly studied Antarctic yeasts Leucosporidium yakuticum AL 101 and Cystobasidium ongulense AL 102 . We have found that LY was a better EPS producer than CO, although both strains possessed an exceptional ability to transform a carbon source into biomass and EPS. The LY-EPS was characterized by a high mannose content, while CO-EPS contained also glucose and galactose. Spectral studies of LY-ЕPS have revealed that mannose units were partly acetylated and at least bound by β-1,4 glycosidic linkages. On the other hand, CO-EPS was more heterogeneous. The EPSs exhibited activation of IFN-γ in splenic F4/80 macrophages and NK cells. We have not got the full picture of the relationship between monosaccharide composition/structure and immunological activity. Taken together, however, our results would seem to suggest that EPS bearing a β-(1,4)-mannan-type structural fragments have a fairly potent activity than more complex and heterogeneous CO-EPS. An important matter to resolve for future studies is the chromatographical separation and purification of EPS into different fractions and the examination of their activity. Research into solving this issue is already underway. Despite this, we believe our work serves as a useful starting point for an assessment of the potential for the application of yeasts and their EPSs.