Volatile Profiling of Pleurotus eryngii and Pleurotus ostreatus Mushrooms Cultivated on Agricultural and Agro-Industrial By-Products

The influence of genetic (species, strain) and environmental (substrate) factors on the volatile profiles of eight strains of Pleurotus eryngii and P. ostreatus mushrooms cultivated on wheat straw or substrates enriched with winery or olive oil by products was investigated by headspace solid-phase microextraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS). Selected samples were additionally roasted. More than 50 compounds were determined in fresh mushroom samples, with P. ostreatus presenting higher concentrations but a lower number of volatile compounds compared to P. eryngii. Roasting resulted in partial elimination of volatiles and the formation of pyrazines, Strecker aldehydes and sulfur compounds. Principal component analysis on the data obtained succeeded to discriminate among raw and cooked mushrooms as well as among Pleurotus species and strains, but not among different cultivation substrates. Ketones, alcohols and toluene were mainly responsible for discriminating among P. ostreatus strains while aldehydes and fatty acid methyl esters contributed more at separating P. eryngii strains.


Introduction
Mushrooms have been considered a delicacy for centuries, due to their distinct texture, aroma and exceptional flavor, while pertinent research evidenced unique nutritional and health-promoting benefits [1][2][3]. Mushrooms' popularity steadily increased during the last decades, leading to a 30-fold increment of the world production of edible mushrooms, with Lentinus edodes and Pleurotus ("oyster") spp. accounting for over 40% of the respective total supply [4]. Pleurotus (oyster) mushrooms are low energy, low fat, low sodium and cholesterol-free food items, being at the same time rich in proteins, minerals, functional polysaccharides like chitin and β-glucans, and water-soluble vitamins. In addition, they contain health-promoting bioactive microconstituents like ergosterol (provitamin D2), phenolic acids, the antioxidant amino acid ergothionein and lovastatin [5,6]. Currently, several species of Pleurotus are produced worldwide (e.g., P. ostreatus, P. pulmonarius, P. eryngii, P. djamor and P. citrinopileatus), by exploiting a wide range of lignocellulosic residues as substrates [7]. Of particular interest is their cultivation on olive mills and wineries by-products (i.e., materials that are difficult to manage due to their high content in harvest, approximately 50 g of fresh mushroom samples from each treatment (in triplicates) were wrapped in aluminum foil, sealed in plastic bags and kept at −40 • C until analysis, which was conducted within one week. Prior to analysis, samples were thawed, finely chopped and approximately 1 g was weighed into 15 mL screw capped glass vials (Supelco, Bellefonte, PA, USA), followed by the addition of 4 mL of saturated aqueous NaCl solution to inhibit enzymatic degradation and boost the release of volatiles [17,29] and 50 µL of internal standard (4-methyl-1-pentanol) methanolic solution (µg/mL). A magnetic stir bar was immersed in each vial, which was sealed with PTFE-faced/silicone septum (Supelco, Bellefonte, PA, USA). Sealed SPME vials were kept in a water bath at 40 • C, under magnetic stirring (250 rpm, Phoenix Instrument RSM-10 HS/HP, Garbsen, Germany) for 5 min to equilibrate. Then the SPME assembly was inserted, and fiber was exposed in the headspace and left to incubate for 40 min under continuous stirring. The volatiles were extracted with 50/30 µm divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) StableFlex fibers (Supelco Inc., Bellefonte, PA, USA), reported exhibiting good selectivity and high efficiency for the extraction of volatiles from wild and cultivated mushrooms [15,17,27,[30][31][32]. They are additionally more selective for the identification of aldehydes such as methional and phenylacetaldehyde, considered important for mushroom flavor [32].

Gas Chromatography-Mass Spectrometry Analysis
An Agilent GC 6890N gas chromatograph (Waldron, Germany) coupled with HP5973 Mass Selective detector (electron impact 70 eV) and split-splitless injector, with a specific 0.75 mm i.d. liner provided by Supelco (Bellefonte, PA, USA) was used for volatiles profiling. Following incubation, the SPME assembly was removed from the vial and inserted into the GC injection port, where it remained for 20 min at 220 • C to desorb volatiles. The separation was achieved in a J&W 122-7032 DB-WAX column (30 m, 0.25 mm i.d., 0.25 µm film thickness). High purity helium was the carrier gas at a constant flow of 1.0 mL/min. GC operated in 1:1 split mode and mass detector operated at full scan mode covering 33-350 m/z mass range. The injector and transfer line temperatures were kept at 220 • C. The oven temperature program was: initial temperature 35 • C for 2 min, followed by a ramp of 20 • C /min to 100 • C, and finally 5 • C /min to 240 • C. The MS data were obtained by the mass detector operating at full scan in the range of m/z 35-350.

Identification and Semi-Quantification of Volatile Compounds
The identification of chromatographic peaks was performed by comparing the mass spectrum of each compound with the Wiley 275 (Wiley, New York, NY, USA) and NIST 98 (NIST MS search v6.1d) mass spectral databases. Further confirmation was performed by calculating the Kovats linear retention indices (RIs), using n-alkanes (C8-C40) standard (Supelco, Bellefonte, PA, USA) as external reference [33] and comparing the values obtained with those reported in the literature [34]. Among the compounds detected, only those exhibiting mass spectra matching qualities higher than 90% and their calculated RI did not differ by more than ±15 from the values available in public domain databases were included in the respective Table 1, Tables S1 and S2. For the aroma description of volatiles, "The Good Scents Company Information System" database was used [35]. Semi-quantification was performed based on the peak area and amount (µg) of the internal standard.

Cooking of Mushroom Samples
Cooking (roasting) of mushrooms was carried out in a domestic oven. For this purpose, approximately 100 g of fresh P. ostreatus LGAM3002 and P. eryngii CS3, cultivated on WS, were weighed, placed in clean Pyrex dishes and roasted at 180 • C for 10 min without aeration. The Pyrex dish was subsequently removed from the oven, covered with aluminum foil and allowed to cool for 5 min. Mushrooms were weighed before and after cooking to estimate water loss. Cooking procedure was conducted in triplicate. The pre-treatment and analysis of roasted samples was the same as that applied for the fresh ones.

Total Lipids Content
Mushrooms lipid content was measured in freeze-dried, pulverized samples by the colorimetric sulfo-phospho-vanillin reaction, employing commercial sunflower oil as lipid standard [36,37]. Table 1. Names, classes, retention times (Rt) and retention indices (RIs) of volatile compounds identified by gas chromatography-mass spectrometry (GC-MS).

No
Class

Statistical Analysis
Analyses were performed in triplicate and data are presented as mean ± standard deviation. Differences between means were established by one way analysis of variance (ANOVA) and Duncan's t-test (p < 0.05) with the SPSS software (SPSS for Windows, version 21.0, SPSS Inc., Chicago, IL, USA). Principal component analysis (PCA) was performed on the entire volatile compounds profile set, as well as on the volatile compounds' classes, to attain an overview of possible associations between volatiles and species/strains, raw and roasted mushroom samples and/or different cultivation substrates. Mean centering was only applied on the respective data sets. R-studio 1.0.136/R3.3.3 loaded with the "ade4" [38] and "adegraphics" [39] packages, were used for PCA.

Total Volatiles Contents
Overall, according to the criteria set at paragraph 2.5, more than 80 compounds were identified in mushrooms of the present study, with 24 of them detected only in roasted samples (Table 1). Fresh P. ostreatus mushrooms contained a lower number of volatile compounds compared to fresh P. eryngii ones (36 vs. 56 compounds). In accordance with our observation, Yin et al. (2019) [40] reported that P. ostreatus had the lower number of volatile compounds among six Pleurotus species. However, one should not overlook that the concentrations of volatile compounds alone are not a sufficient base for studying the aroma profile of mushrooms, as different volatile compounds may present significantly different odor thresholds [30].
Nevertheless, concentrations of the volatiles (i.e., compounds included in Tables 2 and 3) were higher in P. ostreatus compared to P. eryngii (1684 ± 591 vs. 965.4 ± 321.5 µg/g f.w), in agreement with the data reported by Jung et al. (2019) [2]. The compounds detected in fresh samples are classified as aldehydes, alcohols, ketones, alkanes, fatty acid methyl esters (FAME) and terpenes. Moreover, the heterocyclic 2-pentyl-furan and the aromatic hydrocarbons toluene and ethylbenzene were also detected. Regarding the strains studied, P. eryngii LGAM106 and P. ostreatus CS5, CS6 and LGAM3002 presented the higher total volatiles content. In most cases, cultivation on GM and/or OL resulted in enhanced volatiles production however this pattern was not consistent (Figure 1).

Aldehydes
Fifteen aliphatic and aromatic aldehydes were identified in fresh Pleurotus samples, 12 of them in P. ostreatus and 14 in P. eryngii (Tables S1 and S2). Aliphatic aldehydes are biochemically derived from fatty acids, whereas the aromatic ones derive from amino acids [41]. Aldehydes confer a fresh, floral grassy and fatty aroma [17,29], and are characterised by low odor thresholds, i.e., they can be perceived even in small concentrations [40]. As a general trend, three out of four P. eryngii strains contained more volatile aldehydes than P. ostreatus strains, while this was also the case for mushrooms cultivated on WS and/or GM ( Figure 1). Hexanal was the predominant aldehyde in both species (in accordance to the findings of a recent study [2]), followed by 2-octenal, n-octanal and 2-heptenal. P. eryngii contained three times more benzaldehyde than P. ostreatus (17.72 ± 11.17 vs. 5.58 ± 2.96 µg/g f.w.). According to Mau et al. (1998) [41], benzaldehyde was the major volatile compound in fruitbodies of P. eryngii. This aromatic aldehyde derives from the catabolism or oxidative degradation of phenylalanine [29,42,43] which is present in the Pleurotus mushrooms studied [44]. Benzaldehyde together with phenylacetaldehyde and short-chain aldehydes (pentanal, hexanal, octanal, nonanal, trans-2-heptenal) detected in the mushrooms studied (Table 1), are oxidative degradation products of oleic and linoleic acids; according to a recent report, they are among the major contributors of edible oil flavors during ambient storage [45]. Heptanal, 2-hexenal, phenylacetaldehyde, 2,4-nonadienal, and 2,4 decadienals were also present in the volatiles of both mushroom species, at slightly higher concentrations in P. eryngii; finally, nonanal, undecenal and 3-dodecenal were detected only in P. eryngii, while 2-phenyl-2-butenal was recorded only in P. ostreatus. The aldehydes identified in this study have been previously reported in several mushroom species [2,17,29,40].
identified in roasted samples.
Nevertheless, concentrations of the volatiles (i.e., compounds included in Tables 2 and 3) were higher in P. ostreatus compared to P. eryngii (1684 ± 591 vs. 965.4 ± 321.5 μg/g f.w), in agreement with the data reported by Jung et al. (2019) [2]. The compounds detected in fresh samples are classified as aldehydes, alcohols, ketones, alkanes, fatty acid methyl esters (FAME) and terpenes. Moreover, the heterocyclic 2-pentyl-furan and the aromatic hydrocarbons toluene and ethylbenzene were also detected. Regarding the strains studied, P. eryngii LGAM106 and P. ostreatus CS5, CS6 and LGAM3002 presented the higher total volatiles content. In most cases, cultivation on GM and/or OL resulted in enhanced volatiles production however this pattern was not consistent (Figure 1).

Aldehydes
Fifteen aliphatic and aromatic aldehydes were identified in fresh Pleurotus samples, 12 of them in P. ostreatus and 14 in P. eryngii (Tables S1 and S2). Aliphatic aldehydes are biochemically derived from fatty acids, whereas the aromatic ones derive from amino acids [41]. Aldehydes confer a fresh, floral grassy and fatty aroma [17,29], and are characterised by low odor thresholds, i.e., they can be perceived even in small concentrations [40]. As a general trend, three out of four P. eryngii strains contained more volatile aldehydes than P. ostreatus strains, while this was also the case for mushrooms cultivated on WS and/or GM (Figure 1). Hexanal was the predominant aldehyde in both species (in accordance to the findings of a recent study [2]), followed by 2-octenal, n-octanal and 2heptenal. P. eryngii contained three times more benzaldehyde than P. ostreatus (17.72 ± 11.17 vs. 5.58 ± 2.96 μg/g f.w.). According to Mau et al. (1998) [41], benzaldehyde was the major volatile compound in fruitbodies of P. eryngii. This aromatic aldehyde derives from the catabolism or oxidative degradation of phenylalanine [29,42,43] which is present in the Pleurotus mushrooms studied [44]. Benzaldehyde together with phenylacetaldehyde and short-chain aldehydes (pentanal, hexanal, octanal, nonanal, trans-2-heptenal) detected in the mushrooms studied (Table 1), are oxidative degradation products of oleic and linoleic acids; according to a recent report, they are among the major contributors of edible oil flavors during ambient storage [45]. Heptanal, 2-hexenal, phenylacetaldehyde, 2,4-nonadienal, and 2,4 decadienals were also present in the volatiles of both mushroom species, at slightly higher concentrations in P. eryngii; finally, nonanal, undecenal and 3dodecenal were detected only in P. eryngii, while 2-phenyl-2-butenal was recorded only in P. ostreatus. The aldehydes identified in this study have been previously reported in several mushroom species [2,17,29,40].

Fatty Acid Methyl Esters (FAME)
FAME originate from fatty acids metabolism [42,48] and impart fatty and fruity aromas, with very low odor thresholds [17]. In the present study, seven FAME were identified in P. eryngii, and six in P. ostreatus, their total concentrations being 65.79 ± 40.31 vs. 13.91 ± 8.49 µg/g f.w., in line with their respective lipids content (7.29 ± 0.4 vs. 3.73 ± 0.2 mg/g f.w.). Methyl palmitate was the predominant FAME in both species (31.44 ± 20.62 vs. 6.02 ± 3.11 µg/g f.w. in P. eryngii and P. ostreatus) followed by methyl linoleate and methyl oleate (Tables S1 and S2). Methyl pentadecanoate, methyl laurate and methyl myristate were also detected at lower concentrations, while methyl stearate was found only in P. eryngii samples. There is a clear difference regarding the levels of FAME among the species studied since volatiles of P. eryngii strains contain more FAME than P. ostreatus (Figure 1). Regarding the influence of cultivation substrates, volatiles of P. eryngii mushrooms from OL and GM contained more FAME, whereas no clear pattern was evident in P. ostreatus (Figure 1). The FAME identified in this study have been reported in various Pleurotus spp. [51,52], Morchella importuna [17], Agaricus bisporus [27], and eleven wild edible mushrooms [31].

Other Compounds-Toluene
The aromatic hydrocarbons toluene and ethylbenzene as well as the heterocyclic 2-pentylfuran, not belonging to the groups mentioned above, are discussed separately. Toluene was present in all strains of both species at 17.05 ± 8.11 vs. 104.5 ± 38.8 µg/g f.w. in P. eryngii and P. ostreatus, respectively. Toluene derives from unsaturated fatty acids and has been reported in wild edible species [15], freeze-dried truffle (Tuber aestivum) [57], and other food items like cooked beef [58,59] and fresh Mediterranean fish and shellfish [60]. Toluene exhibited a distinct distribution among Pleurotus volatiles, its concentrations being 3-13 times lower in P. eryngii when compared to P. ostreatus (Figure 1, Tables S1 and S2). Ethylbenzene was detected at low concentrations only in P. eryngii. It has been reported in food items like cooked beef [58] and fermented cucumbers [61] but-to the best of our knowledge-not in edible mushrooms. The heterocyclic compound 2-pentylfuran is formed by the autoxidation of linoleic acid [62]; its odor is described as green, earthy and meaty [17,40]. It was detected at low concentrations in P. eryngii and P. ostreatus (0.17-0.94 µg/g f.w.), and it did not exhibit any specific distribution pattern. It was identified in Pleurotus [40] and other edible mushrooms [2,15,17,50].

Eight-Carbon Compounds
Eight-carbon volatile compounds are major contributors to the characteristic flavor of many mushrooms [2,17,21,40,49]. They are oxidation products of linoleic acid via a 10-hydroperoxide intermediate [21,23,40] which represents 60.6-80.6% of fatty acids in P. eryngii and P. ostreatus [3,7]. They are classified as oxylipins and are involved in a wide range of biological processes [21]. Eight-carbon compounds predominated among the volatiles of all Pleurotus strains studied, comprising 78-83% and 84-91% of total volatiles in P. eryngii and P. ostreatus, following similar distributions with the total volatiles presented in Figure 1. Their concentrations were 776.5 ± 282.0 µg/g f.w. in P. eryngii and almost doubled in P. ostreatus (1498 ± 551 µg/g f.w.) ( Table 2). In addition, 1-octen-3-ol predominated in all P. eryngii and in three out of four P. ostreatus strains, comprising 88-95% and 47-53% of eight-carbon compounds, respectively. Differences were observed among the Pleurotus strains studied: (a) three of the eight-carbon compounds detected in P. eryngii were absent from P. ostreatus, and (b) in three out of four P. eryngii strains, the second more abundant eight-carbon compound was 3-octanone followed by 2-octenal, while in three out of four P. ostreatus strains it was 3-octanol, followed by 3-octanone (Table 2).

Effect of Cooking on Volatiles' Profiles
The volatile compounds determined in fresh and roasted mushrooms of P. eryngii strain CS3 and P. ostreatus strain LGAM 3002 cultivated on wheat straw are shown in Table 3. Besides water loss, roasting caused significant alterations in the volatiles' profiles for both species. Overall, the number of volatile compounds decreased from 35 to 27 in roasted P. eryngii and increased from 35 to 41 in roasted P. ostreatus (Table 3). The percent abundance of volatile alcohols decreased by 65% and 77% in roasted P. eryngii and P. ostreatus, respectively; the same trend was followed by fatty acid methyl esters, which were almost eliminated (they decreased by 95% and 97% in roasted P. eryngii and P. ostreatus, respectively). Similarly, ketones, aldehydes and eight-carbon compounds decreased by 66%, 64% and 65% in P. eryngii and by 85%, 9.0% and 81% in P. ostreatus. Elimination of FAME during the freeze-drying of A. bisporus has been reported by Pei et al. (2016) [27].
In addition, roasting resulted in the formation of nitrogen-containing heterocyclic compounds (17 pyrazines, one pyrrole and one oxazole) and two sulfur compounds, i.e., substances with low odor threshold values, known to enrich food flavor with meat-or roast-like aromas [63]. The predominant compounds in roasted P. eryngii were 1-octen-3-ol (mushroom-like flavor), followed by hexanal (green and woody), toluene (sweet, solventy, woody, roasted coffee) and undecane (gasoline-like). In roasted P. ostreatus 1-octen-3ol also predominated, followed by 3-octanol (mushroom, buttery), toluene, 3-octanone (mushroom, ketonic, cheesy) and 3-ethyl-2,5-dimethyl pyrazine (hazelnut, nutty). Several aroma compounds detected in fresh and cooked mushrooms originate from the enzymatic and oxidative decomposition of unsaturated fatty acids or from reactions between amino acids and carbonyl compounds, i.e., the Maillard reactions and the Strecker amino acid degradation which leads to the formation of Strecker aldehydes and other flavor-active compounds [48,64,65]. The extent of these transformations depends on the types of compounds involved, temperature, pH and reaction time [65][66][67]. Regarding the influence of temperature, the volatile compounds produced by lipid oxidation (aldehydes, ketones, and aromatic hydrocarbons like hexanal, octanal, benzaldehyde and toluene) are formed at relatively low temperatures, even at room temperature [45]; this provides an explanation for their presence in the fresh mushroom samples (Table 3). By contrast, Maillard and Strecker reactions take place at higher temperatures. Literature data consider temperatures around 120-130 • C as optimal for the formation of Maillard reaction products with favorable sensory characteristics [68] as well as for the formation of compounds with sulfur and nitrogen groups [69,70], e.g., pyrazines, Strecker aldehydes, dimethyldisulfide and pyrrole (detected in the roasted Pleurotus samples; Table 3). It is noteworthy that Pleurotus mushrooms contain significant amounts of Maillard reaction precursors, namely free amino acids, reported to comprise 37.1-41.6% of crude protein [44] and glucans, which comprise 40-50% w/w of the total content on a dry weight basis [71]. In addition, the Strecker aldehydes 2-methylbutanal, 3-methylbutanal, methional, phenylacetaldehyde and benzaldehyde, detected in the roasted samples (Table 3) derive from isoleucine, leucine, methionine and phenylalanine [67,72], i.e., amino acids that are present in the mushrooms studied [44].
It is noteworthy that due to the absence of thermal treatments, pyrazines, pyrrole, oxazole and sulfur compounds were detected only in the cooked samples (Table 3); hence, the volatile profiles of fresh mushrooms obtained in the present study can be considered as representative, in some degree, of the P. eryngii and P. ostreatus aroma.
The sulfur compounds methional and dimethyl disulfide were detected in the roasted Pleurotus samples (Table 3). Methional is the Strecker aldehyde of methionine and is eventually decomposed to form dimethyl disulfide, which contributes further to the overall flavor development [65,67]. Methional is responsible for the aroma of cooked potatoes [67] and is the key aroma-active compound in cooked T. matsutake mushrooms [55], while both sulfur compounds were reported in A. bisporus soup [82].

Principal Component Analysis (PCA)
Principal component analysis (PCA) was performed on the entire volatile compounds profile of P. eryngii and P. ostreatus strains, in order to investigate the existence of any groupings or associations (Figure 2). The first two principal components, explaining 98% of the data set variance (i.e., PC1: 75.4% and PC2: 22.6%), allowed a clear discrimination of two Pleurotus species and eight strains studied (Figure 2a,b; ellipses drawn at a confidence level of 0.95). The sulfur compounds methional and dimethyl disulfide were detected in the roasted Pleurotus samples (Table 3). Methional is the Strecker aldehyde of methionine and is eventually decomposed to form dimethyl disulfide, which contributes further to the overall flavor development [65,67]. Methional is responsible for the aroma of cooked potatoes [67] and is the key aroma-active compound in cooked T. matsutake mushrooms [55], while both sulfur compounds were reported in A. bisporus soup [82].

Principal Component Analysis (PCA)
Principal component analysis (PCA) was performed on the entire volatile compounds profile of P. eryngii and P. ostreatus strains, in order to investigate the existence of any groupings or associations (Figure 2). The first two principal components, explaining 98% of the data set variance (i.e., PC1: 75.4% and PC2: 22.6%), allowed a clear discrimination of two Pleurotus species and eight strains studied (Figure 2a,b; ellipses drawn at a confidence level of 0.95). However, due to the high number of identified compounds, it was impossible to acquire some additional information regarding the volatile compounds that were responsible for the species/strain discrimination through the interpretation of their loadings. Therefore, a PCA was performed on the basis of the volatile compounds' classes, which revealed a clear separation at species and strain level (Figure 3a,b, respectively). Furthermore, the first two principal components explained an even higher percentage (98%) of the data set variance (i.e., PC1: 95.7% and PC2: 2.9%). Separation is taking place in both axes (PC1 and PC2), with P. eryngii strains mainly placed in the 4th quartile and P. ostreatus strains in the 2nd quartile. The observation of the loadings (i.e., the compound classes) revealed a positive effect of ketones, alcohols and toluene at separating P. ostreatus strains (Figure 3c). On the other hand, aldehydes and fatty acid methyl esters presented a stronger effect on the separation of P. eryngii strains. Both observations are in agreement with the previously determined content in volatile compounds for these particular species/strains. However, due to the high number of identified compounds, it was impossible to acquire some additional information regarding the volatile compounds that were responsible for the species/strain discrimination through the interpretation of their loadings. Therefore, a PCA was performed on the basis of the volatile compounds' classes, which revealed a clear separation at species and strain level (Figure 3a,b, respectively). Furthermore, the first two principal components explained an even higher percentage (98%) of the data set variance (i.e., PC1: 95.7% and PC2: 2.9%). Separation is taking place in both axes (PC1 and PC2), with P. eryngii strains mainly placed in the 4th quartile and P. ostreatus strains in the 2nd quartile. The observation of the loadings (i.e., the compound classes) revealed a positive effect of ketones, alcohols and toluene at separating P. ostreatus strains (Figure 3c). On the other hand, aldehydes and fatty acid methyl esters presented a stronger effect on the separation of P. eryngii strains. Both observations are in agreement with the previously determined content in volatile compounds for these particular species/strains. PCA was also performed on the basis of volatile compounds' classes for cooked (roasted) and raw (fresh) P. eryngii strain CS3 and P. ostreatus strain LGAM 3002 cultivated on wheat straw, to detect any groupings and to acquire additional information. The PCA allowed a clear discrimination of samples both in terms of species/strains as well as of the roasted/raw treatment (Figure 4a,b). As regards the latter, in particular, separation took place across PC1 (X-axis) with the fresh samples being placed on the negative side of PC1 and the roasted on the positive side. By examining the PCA loadings, it appears that pyrazines, sulfur compounds and toluene exert a significant role at separating the roasted samples. However, PCA was not able to discriminate mushrooms according to the cultivation substrate on the basis of the volatile compounds or compound classes studied (data not shown). This possibly indicates a low effect of the cultivation substrates on the mushrooms content in volatile compounds, contrary to the outcomes of our previous works regarding antioxidant properties and bioactive microconstituents like ergosterol, phenolic and terpenic compounds as well as free amino acids in the same mushroom species [5,44]. PCA was also performed on the basis of volatile compounds' classes for cooked (roasted) and raw (fresh) P. eryngii strain CS3 and P. ostreatus strain LGAM 3002 cultivated on wheat straw, to detect any groupings and to acquire additional information. The PCA allowed a clear discrimination of samples both in terms of species/strains as well as of the roasted/raw treatment (Figure 4a,b). As regards the latter, in particular, separation took place across PC1 (X-axis) with the fresh samples being placed on the negative side of PC1 razines, sulfur compounds and toluene exert a significant role at separating the roasted samples. However, PCA was not able to discriminate mushrooms according to the cultivation substrate on the basis of the volatile compounds or compound classes studied (data not shown). This possibly indicates a low effect of the cultivation substrates on the mushrooms content in volatile compounds, contrary to the outcomes of our previous works regarding antioxidant properties and bioactive microconstituents like ergosterol, phenolic and terpenic compounds as well as free amino acids in the same mushroom species [5,44].

Conclusions
The aroma of Pleurotus mushrooms is formed by several classes of compounds namely alcohols, aldehydes, ketones, FAME, alkanes and terpenes. Fifty volatile compounds were identified in P. eryngii strains, while only 31 in P. ostreatus. However, in most cases, P. ostreatus presented a higher content in volatile alcohols, ketones and toluene compared to P. eryngii, whereas the opposite was established for aldehydes and FAME. Roasting of mushrooms caused partial elimination of existing volatiles and the formation of Maillard reaction and Strecker degradation products, mainly pyrazines, which contribute to the distinctive cooked-mushrooms aroma. Unsaturated alcohols and ketones containing eight carbon atoms predominated among the volatiles of both raw and cooked mushrooms. Principal component analysis performed on the aroma profiles data obtained succeeded in discriminating among Pleurotus species/strains and roasted/fresh samples, but not among mushrooms deriving from different cultivation substrates. The findings from the present work are interesting considering that previous studies on the same mushrooms revealed an influence of substrates on the antioxidant activity, β-glucans content and the levels of bioactive microconstituents like ergosterol, phenolic and terpenic acids, and free amino acids [5,45]. The extent to which the mushrooms organoleptic properties are affected by the choice of species/strains or the type of substrate needs to be further investigated by including the determination of additional compounds and/or sensory evaluation.