Metabolic Profiles and Free Radical Scavenging Activity of Cordyceps bassiana Fruiting Bodies According to Developmental Stage

The metabolic profiles of Cordyceps bassiana according to fruiting body developmental stage were investigated using gas chromatography-mass spectrometry. We were able to detect 62 metabolites, including 48 metabolites from 70% methanol extracts and 14 metabolites from 100% n-hexane extracts. These metabolites were classified as alcohols, amino acids, organic acids, phosphoric acids, purine nucleosides and bases, sugars, saturated fatty acids, unsaturated fatty acids, or fatty amides. Significant changes in metabolite levels were found according to developmental stage. Relative levels of amino acids, purine nucleosides, and sugars were higher in development stage 3 than in the other stages. Among the amino acids, valine, isoleucine, lysine, histidine, glutamine, and aspartic acid, which are associated with ABC transporters and aminoacyl-tRNA biosynthesis, also showed higher levels in stage 3 samples. The free radical scavenging activities, which were significantly higher in stage 3 than in the other stages, showed a positive correlation with purine nucleoside metabolites such as adenosine, guanosine, and inosine. These results not only show metabolic profiles, but also suggest the metabolic pathways associated with fruiting body development stages in cultivated C. bassiana.

A metabolomic approach has been used for major metabolite identification in fungus as well as for fungal taxonomy of Penicillium using mass spectrometry [17,18] and for characterization of secondary metabolites of Aspergillus [19].Only one metabolic profiling study on C. bassiana during development using nuclear magnetic resonance spectrometry has recently been reported [20].However, the investigation of the major metabolites and metabolism associated with and the free-radical scavenging activities of cultivated C. bassiana have not yet been elucidated. Moreover, most studies on Cordyceps species during fruiting body formation have only focused on targeted chemical composition change [21] and gene expression profiling [22,23].
The enrichment analysis method was recently developed for the functional interpretation of large amounts of data in the fields of genomics, transcriptomics, proteomics, and metabolomics [24,25]. Enrichment analysis is a useful tool to investigate wide ranges of biological and chemical annotations in several organisms [26]. Recently, studies on biomarker annotation in human liver tissue, metabolomic correlation networks in Arabidopsis, and a global test for metabolic pathways in Escherichia coli and Saccharomyces cerevisiae under different conditions have been reported by functional enrichment analysis [27][28][29].
However, to the best of our knowledge, no research using enrichment analysis has investigated metabolite annotation or metabolism changes in C. bassiana according to fruiting bodies, which are classified from stages 1 to 4 during the formation of stromata and perithecium of fruiting body surfaces (stage 1,prior to perithecium formations; stage 2,early perithecium formation; stage 3,completed perithecium formation; and stage 4,aging after perithecium formation). Perithecia that from the stromata are flask-shaped structures containing ascospores, and the characteristic morphology of C. bassiana stromata has been reported previously [8].Thus, the developmental stages of C. bassiana fruiting bodies were categorized by the degree of perithecium formation in this study.
The main hypothesis is that the metabolite level associated with specific metabolisms and free-radical scavenging activity might change according to developmental stage of cultivated C. bassiana.
In this study, we performed metabolic profiling of cultivated C. bassiana in various developmental stages using gas chromatography-mass spectrometry (GC-MS). In addition, the free-radical scavenging activities of those samples and their correlation with specific metabolites were investigated. The main objectives of this study were metabolic profiling and investigation of the free-radical scavenging activities in cultivated C. bassiana at various developmental stages. The major metabolic pathways associated with developmental stages will also be discussed.

Sample preparation of fruiting body
The anamorph of C. bassiana is B. bassiana, and the nuclear intergenic region Bloc and the nuclear ribosomal internal transcribed spacer region (ITS) were used to confirm the isolates used for fruiting body development in this study were C. bassiana based on the recent phylogenetic analyses of Rehner et al [30].For the artificial production of fruiting bodies, C. bassiana strains were grown on Sabouraud dextrose +1% (w/v) yeast extract broth (SDY) for 3 days at 25uC as inocula for the production of fruiting bodies of C. bassiana. The cultures of fruiting body were grown in 1,000-ml plastic bottles containing brown rice medium and incubated at 20uC under light intensity of 400 lux and 90% humidity. Samples were collected every week from the fourth to seventh weeks of cultivation to monitor and compare the development of stromata and perithecia. Observations were made using a dissecting microscope (SZ2-1ILST, Olympus, Japan) and the four developmental stages were classified as follows: stage 1, week 4 (no perithecia formation on any stroma); stage 2, week 5 (initiation of perithecium formation, yellowish tiny spots appear on the stromata); stage 3, week 6 (club-shaped perithecium formation, filiform ascospores produced in the perithecium); stage 4, week 7 (aging stage, ascospores released from the perithecia). The fruiting bodies isolated from culture media were freeze-dried and powdered according to developmental stage. Samples were stored at 270uC before analysis. Voucher specimens were deposited at the College of Pharmacy, Chung-Ang University, Republic of Korea (fruiting bodies: CAUCBF 20110904-20110907).

Sample preparation for GC-MS analysis
Methanol and n-hexane were used as solvents for the extraction of polar and nonpolar metabolites, respectively. Twenty milligrams of each sample in different growth conditions were transferred into glass Eppendorf tubes (Axygen, Union City, CA) and extracted with 1 mL of 70% methanol and 100% n-hexane. After sonication, the tube was centrifuged at 2,000 rpm for 10 min. The supernatant was collected separately from each sample and filtered through a 0.45-um filter (Acrodisc Syringe Filters, Pall Corporation, NY). After extraction, 100 mL of each sample solution was separately transferred into GC vials and then dried with nitrogen gas flow for 5 min at 60uC. After drying, 30 mL of 20,000 mg/mL methoxylamine hydrochloride in pyridine was added. Thereafter, 50 mL of BSTFA (N,O-bis (trimethylsilyl) trifluoroacetamide; Alfa Aesar, Ward Hill, Massachusetts) containing 1% TMCS (trimethyl chlorosilane) and 10 mL of 2chloronaphthalene (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan; 250 mg/mL in pyridine as an internal standard) were added to the dried sample. Derivatized samples were incubated at 60uC for 60 min, after which the solutions were directly used for GC-MS analysis.For quantification of purines (adenosine, guanosine, inosine, uric acid) in C. bassiana samples, standard solutions (1-100 mg/mL) and sample solution (10,000 mg/L) were prepared with 70% methanol. The sample and each standard solution of 90 mL were transferred into GC vial that was dried with nitrogen gas for 5 min at 60uC.The derivatization was performed as above described method. After derivatization process, the solution was directly used for GC-MS analysis. Characteristic ions of adenosine (230 m/z), guanosine (324 m/z), inosine (217 m/z), and uric acid (411 m/z) were selected in preliminary analysis, and those were used for each purine quantification of C. bassiana fruiting bodies.

GC-MS analysis
Samples were analyzed using a model 7890A Agilent GC (Agilent Technologies, CA) equipped with a model 5975C MSD detector (Agilent Technologies), an autosampler (7683 B series, Agilent Technologies), a split/splitless injector, an injection module, and Chemstation software. The GC inlet temperature was set to 250uC with an injection volume of 1.0 mL and a split ratio of 1:10, using helium as a carrier gas in constant-flow mode of 1.0 mL/min. A fused silica capillary column of 5% phenyl methylpolysiloxane phase (DB-5, Agilent Technologies) with dimensions 30 m60.25 mm i.d. 60.25 mm film thickness was used for analysis. The detector voltage was set to 1518 V, the auxiliary temperature was set to 280uC, the MS source temperature was set to 230uC, and the MS quad temperature was set to 150uC. The mass range was 50-700 Da. Data were obtained in full scan mode. The oven temperature for polar metabolite analysis was 80uC (hold 3 min) programmed to 130uC (3uC/min) then to 240uC (5uC/min) then to 320uC (10uC/min; hold 3 min). For the non-polar metabolite analysis, the detector voltage was set to 1588 V, and the mass range was 50-600 Da. The oven temperature was 80uC programmed to 260uC (5uC/min) then to 300uC (3uC/min; hold 3 min).

Data analysis and enrichment analysis
Raw GC-MS data were processed as described by Styczynski et al. [31]to quantitatively compare global metabolites among all samples. Initially, the AMDIS (Automated Mass Spectral Deconvolution and Identification System, http://chemdata.nist. gov/mass-spc/amdis/) was used for mass spectral deconvolution, which separated peaks from noise and overlapping peaks. Then, the ELU files were subsequently analyzed with an online peakfiltering algorithm (SpectConnect, http://spectconnect.mit.edu). The identification was performed using spectra of individual components transferred to the NIST mass spectral search programs MS Search 2.0, where they were matched against the NIST MS library; for identification, a match quality of 70% was generally accepted. Peak areas of multiple derivative peaks belonging to one compound were summed and considered as a single compound. Normalization to an internal standard peak area was performed before multivariate statistical analyses.
The relative intensities of assigned metabolites by GC-MS analysis were analyzed in each sample. Significant differences in metabolite levels were detected by one-way analysis of variance (ANOVA) using PASW Statistics 18 software (IBM, Somers, NY) followed by Tukey's significant-difference test. The level of statistical significance was set at p,0.05. Partial least-squares discriminant analysis (PLS-DA) was performed in SIMCA-P software (version 12.0, Umetrics, Umeå, Sweden) using meancentered and unit variance-scaled data, which yields a clearer differentiation of each class and enables a less complicated investigation of marker compounds than principal component analysis by rendering the class to each sample group [32].
Measures of model quality were reported for PLS-DA; the cumulative values of total Y explained variance (R 2 ), which describes the goodness of fit, and Y predictable variation (Q 2 ), which measures the predictive power of the model. A data functional enrichment analysis with identified metabolic data for biological information was performed using MBRole [33], freely available from http://csbg.cnb.csic.es/mbrole/. The interface used ID conversion utility, which performs the analysis using the metabolites identified by KEGG (Kyoto Encyclopedia of Genes and Genomes; http://www.genome.jp/kegg/). We selected Aspergillus niger as a background set. The result contains the list annotation over-represented in the input set with respect to the background set and metabolite-associated p-values.

Free radical scavenging activity
The freeze-dried C. bassiana (20 g) grown to different stages was extracted in screwcap vials with 400 mL of 70% methanol. The samples were irradiated four times in a microwave irradiation machine (MARSX, CEM Corporation, NC) for 10 min at 80uC. After extraction, the samples were filtered with filter paper (Whatman No. 4, Whatman, Kent, UK), freeze-dried (FDU-1200, EYELA, Miyagi, Japan) for 48 hours and then stored at 280uC for antioxidant activity analysis. The free radical scavenging ability of C. bassiana was determined by following the procedures by Kovatcheva-Apostolova et al. [34] with some modifications. The microwave extract sample solutions (10,000 mg/L) of 0.2 mL were mixed with 3.8 mL of 6610 25 mM 2,2-diphenyl-1-picryl hydrazyl (DPPH) (Sigma, St. Louis, MO) solution. The mixture was incubated for 30 min in the dark at room temperature. The antioxidant activity of the plant was measured at 515 nm with a microplate spectrophotometer (xMark, Biorad, Berkeley, CA). Free radical scavenging activity (%) was calculated by the following formula: DPPH radical scavenging activity (%)1 A control is the absorbance without sample, and A sample is the absorbance with sample.

PLS-DA
PLS-DA was performed to compare the metabolic profiles of fruiting bodies at different developmental stages. The PLS-DA model was cross-validated. The goodness of fit and predictive ability of the PLS-DA model were quantified using R 2 Y, while the predictive ability of the model was indicated by Q 2 Y,respectively [35].The model performance parameters listed in Table 3;the results strongly suggest that the original models were valid. Both PLS-DA models had appropriate R 2 Y intercept values of 0.242 and 0.164, and Q 2 Y intercept values of 20.394, and 20.444, respectively. In general, models with an R 2 Y intercept of less than 0.4 and a Q 2 Y intercept of less than 0.05 are considered to be valid [32].
The metabolic profile of the 70% methanol extracts of fruiting bodies at each developmental stage is represented in the PLS-DAderived score plot as a single point (R 2 = 0.96, Q 2 = 0.89) in Figure 1A. Stage 3 samples were clearly separated from those samples of stages 1, 2, and 4 by principal component 1, which explained 72.9% of the variance. The samples of stages 1, 2, and 4 were separated mainly by principal component 2. Metabolic profiling of the PLS-DA score plot yielded similar findings to those  obtained using the NMR data, but with different identified metabolites [20].As shown in Figure 1B, the samples from each condition could also be well separated in PLS-DA-derived score plots for 100% n-hexane extracts (R 2 = 0.95, Q 2 = 0.91). Samples from stage 4 were clearly separated from the other sample groups, mainly by PLS component 1, which indicates that profiles of nonpolar metabolites in stage 4 samples were clearly distinguished from the samples at other stages.
Metabolic pathway analysis using enrichment analysis and ANOVA.
The results of the ANOVA conducted to compare differences in the relative levels of the metabolites at various developmental stages are presented in Tables S1 and S2. As shown in Fig. 2, significant changes in metabolites associated with sugar metabolism, purine metabolism, and amino acid metabolism were observed at stage 3. Most of the metabolites, threonine, serine, valine, fructose, arabinose, ribose, isoleucine, lysine, glycine, histidine, glutamine, glucose, and aspartic acid, were associated with ABC transporters and were present at a higher level in the completed perithecium formation stage (stage 3) than in the other stages. ABC-transporter proteins such as PDR, MDR, and MRP have been found in Saccharomyces cerevisiae and Schizosaccharomyces pombe [36]. It has been reported that the antioxidant capability in B. bassiana is reduced when ABC-transporter genes such as Pdr, Mdr, and Mrp are disrupted by mutation [37]. It is therefore expected that antioxidant activity will be higher at stage 3, in which higher levels of metabolites associated with ABC transporters were observed.
Energy-consuming ABC transporters exist in microbial cell membranes. These transporters play important roles in regulating nutrient uptake and secreting toxins or antimicrobial agents [38][39][40].It seems that various metabolites related to ABC transporters are needed and highly accumulate during stage 3 in order to regulate nutrient uptake or toxin secretion for ascospore formation. Besides ABC-transporters genes, the many transporters have been studied in various fungal species [41][42][43] However, to our knowledge the present report is the first regarding annotation of the metabolites related to ABC transporters in C. bassiana fruiting bodies at different development stages. Further studies on the roles of various transporters at different developmental stages of C. bassiana are needed.
Higher levels of amino acids associated with purine metabolism, such as glutamine and glycine, were observed at stage 3. Not only are the levels of amino acids such as tyrosine, tryptophan, asparagine and GABA increased at stage 3, but so also are those of purines such as adenosine, inosine, guanosine, and uric acid. Therefore, amino acid metabolism and purine metabolism may play important roles in regulating the development of cultivated C.bassiana fruiting bodies.
The levels of sugars such as fructose, galactose, glucose, and mannose were significantly increased at stage 3 (Fig. 2). A carbon source, such as glucose, is reported to act as an intermediate of glycolysis, the TCA cycle, and in the metabolism of purine, alanine, aspartate, and glutamate [44].Higher levels of the TCAcycle intermediates aconitate and malate were also observed at stage 3, whereas the levels of citrate, succinate, and fumarate were lower. As indicated in Figure 2 and Table S2, levels of saturated fatty acids such as arachidic acid, behenic acid, and lignoceric acid were higher levels in the no-perithecium (stage 1) and early perithecium formation (stage 2) stages, whereas both saturated and unsaturated fatty acids were decreased in the aging stage (stage 4). Upregulated lipid biosynthesis gene expression has been reported during the early stages of perithecium development, together with a reduction in the levels and types of fatty acids [45]. Various lipids including linoleic acid were found to be associated with perithecium development in Nectria haematococca [46]. The lower levels of fatty acids observed in the present study in the cultivated C. bassiana at stage 4 may be attributable to the enhanced usage of the fatty acids as energy resources for ascospores release from the perithecia.
Among the various fatty acids, the levels of palmitic and linoleic acids which are known key intermediates involved in membrane biogenesis were higher level at stage 3 than at the other stages. It appears that palmitic and linoleic acids accumulate during stage 3 in preparation for ascospore formation.
Higher levels of putrescine and nicotinic acid, (also known as vitamin B 3 ), were observed at stage 4. Putrescine is a breakdown product of amino acids [47] and is involved in the reduction in amino acids observed in stage 4 samples. It is interesting to note that the levels of most of the fatty acids had decreased at stage 4, whereas that of nicotinic acid increased. It may be that the fatty- acid breakdown products are converted into putrescine and nicotinic acid during fruiting body formation.
To summarize, it appears that various metabolites are needed for the formation of ascospore in the perithecium in stage 3. Transcriptomic and proteomic approaches will be applied in the near future to investigate or confirm the mechanism underlying the variations in metabolic profiles that occur during the various development stages of the C. bassiana fruiting body.

DPPH radical-scavenging activity
Variations in the DPPH radical-scavenging activities during the development of C. bassiana fruiting bodies were investigated; the results are summarized in Table 5. The free-radical-scavenging activities of stages 1, 2, and 4 were similar, whereas those of stage 3 were significantly higher. Ascorbic acid was used as a positive control. For antioxidant activity, the DPPH radical-scavenging  activity of stage 3 (10,000 mg/L) was 47.7%, while that of ascorbic acid (50 mg/L) was 41.1%.In C. bassiana, the total phenol content was higher in stage 3 samples (i.e., after 6 weeks of cultivation) than in samples from other stages [48]. It has been reported that the activities of superoxide dismutase and catalase in B. bassiana prevent an increase in reactive oxygen species and the subsequent damage that they can cause [14].Purine ribonucleosides, including guanosine, inosine, and adenosine, have been reported to protect against DNA oxidative damage by reducing the production of hydrogen peroxide and hydroxyl radicals [49].Furthermore, uric acid, which scavenges hydroxyl radicals, is known to be an important antioxidant [50]. Adenosine, guanosine, inosine, and uric acid, the levels of which are enhanced at stage 3, are associated with antioxidant activity. As shown in Figure 3, strong positive correlations (coefficients r 2 = 0.99) were found in the present study between the content of purine nucleoside compounds (Table S4) and antioxidant activities ( Table 5). Adenosine is also known to act as an antiwrinkle compound for human skin [51,52].Thus, cultivated fruit bodies of C. bassiana at stage 3 might be considered natural resources for nutraceuticals with free-radical scavenging activity or as cosmeceutical materials with antiwrinkle activity.

Conclusions
In this study, we performed metabolic profiling of C. bassiana using a GC-MS-based non-targeted profiling approach. Alteration of several major metabolic pathways, including sugar metabolism, purine metabolisms, amino acid metabolism, TCA cycle, and lipid metabolism, were observed during fruiting body development. Especially, in the perithecium formation stage (stage 3), the relative levels of metabolites associated with ABC transporters, aminoacyl-tRNA biosynthesis, and purine metabolism were significantly increased, whereas levels of metabolites associated with lipid metabolism decreased in the ascospores released stage (stage 4).The free radical scavenging activity, which was significantly higher in stage 3 than in the other stages, was positively correlated with purine derivatives such as adenosine, guanosine, inosine and uric acid. Thus, C. bassiana of stage 3 was suggested to be a better resource with higher level of free-radical scavenging activities compared to the other fruiting body samples of C. bassiana at other stages. We suggest that the metabolic profiles during fruiting body development determined in this study provide useful criteria for selecting appropriate harvest points of fruiting bodies of C. bassiana. To our knowledge, this is the first study to investigate the correlation between metabolic profiles and free radical scavenging activities of C. bassiana fruiting bodies according to developmental stage. Figure S1 Representative GC-MS spectra of 70% methanol (A), and 100% n-hexane (B) extracts of cultivated C. bassiana at various development stages.

(TIF)
Table S1 Metabolites identified by GC-MS analysis of 70% methanol extracts in C. bassiana fruiting bodies. The relative levels of each metabolite were obtained by dividing the area % of metabolite by the area % of internal standard. Different letters in the same row represent a significant difference. Data are mean 6 STD values for triplicate measurements. ND, not detected in the sample.