Volatile Organic Compounds from Filamentous Fungi : a Chemotaxonomic Tool of the Botryosphaeriaceae Family

Volatile organic compounds (VOCs) from ten endophytic fungal species belonging to the Botryosphaeriaceae family were extracted by headspace-solid phase micro-extraction (HS-SPME) and analyzed by gas chromatography-mass spectrometry (GC-MS). Thirty-four VOCs were identified. Most of the compounds are sesquiterpenes (14 non-oxygenated and 10 oxygenated), and two linear ketones and eight alcohols were also identified. Multivariate data analysis (PCA and HCA) allowed the differentiation of all investigated species, and proved to be efficient for the differentiation of Neofusicocum parvum and N. ribis, which are considered very similar species. α-Bisabolol, α-selinene, α-cedrene epoxide and guaiol acetate were suggested as biomarkers.


Introduction
Botryosphaeriaceae comprises a broad group of cosmopolitan filamentous fungi (endophytes, pathogens or saprobes) that inhabits several hosts. 1,2Although endophytic fungi are viewed as plant mutualists, under certain circumstances these microorganisms can become phytopathogens. 3This phenomenon has been reported in some species of Botryosphaeriaceae, which have caused disease in agronomically important plants. 2,4Lasiodiplodia theobromae, Neofusicoccum luteum, N. parvum, N. australe, Botryosphaeria dothidea, Diplodia mutila, D. seriata, Dothiorella iberica and D. viticola are examples of phytopathogenic Botryosphaeriaceae fungi that cause damage to Vitis vinifera, 5 Eucalyptus globulus, 4 E. urophylla 6 and Mangifera indica. 7he taxonomy of Botryosphaeriaceae is rather confusing, mainly because of the overlay of the morphological characteristics and species diversity.Several anamorphic species are described for this family belonging to several genera such as Botryodlplodla, Diplodia, Dothiorella, Fusicoccum, Lasiodiplodia, Macrophoma and Sphaeropsis. 8rous et al. 1 reported the existence of anamorphic forms of Botryosphaeria species, which present morphological characteristics of both Diplodia and Fusicoccum genera.Data from DNA (28S rDNA) sequencing were used to differentiate ten phylogenetic strains and to group some of them of the Botryosphaeriaceae family.According to Phillips et al., 2 studies on morphological characters are inadequate to define or identify Botryosphaeriaceae species, and taxa with no DNA sequencing data should not be grouped in this family.Although phylogenetic studies using data from molecular biology have contributed significantly to the taxonomy of Botryosphaeriaceae, 1,2,9 some problems still exist.For instance, Neofusicocum parvum and N. ribis, classified as a complex N. parvum/N.][12] About 10,000 microbial species have been described in the literature, although the microbial volatile organic compound (mVOC) of only a reduced number of them (349 bacteria and 69 fungi species) have been investigated. 13][15] Larsen and Frisvad 16 were the first to demonstrate the use of mVOC for the discrimination of Penicillium species.Strains of dermatophytic fungi were also differentiated based on mVOC, suggesting the use of this analytical tool in early diagnosis and treatment of contaminated patients. 17Volatile profiles of nine root-associated fungal strains (eight species) from three different functional groups (ectomycorrhizal, pathogenic and saprophytic) were successfully used as a chemotyping tool for non-invasive identification of these microorganisms. 18mong the different techniques used to obtain volatile and semi-volatile compounds (terpenes and other classes of VOC), the solid phase micro-extraction (SPME) stands out for its practicality in the sample preparation and analyte preconcentration, under relatively mild conditions. 4,19Studies on volatile metabolites from fungi have involved the use of headspace-solid phase micro-extraction (HS-SPME), by having polydimethylsiloxane/divinylbenzene (PDMS/DBV) as the most efficient fiber mixture. 19,20][20][21] Principal component analysis (PCA) and hierarchical cluster analysis (HCA) are among the statistical techniques (multivariate data analysis) most used for the analysis of VOCs obtained by HS-SPME associated with GC-MS.For example, three chemotypes of Lippia graveolens HBK, distributed in eight populations from Guatemala, were differentiated by these two techniques. 22n this work we describe the use of HS-SPME followed by GC-MS analysis to study VOCs produced by ten species of endophytic fungi from the Botryosphaeriaceae family associated with plants from the Caatinga biome (state of Ceará, Brazil).In addition, the multivariate data analyses PCA and HCA were used to establish differentiation patterns of the investigated species, and to identify biomarkers for the chemotaxonomic classification of these species.

Cultivation of fungi, extraction and analysis of the volatile organic compounds, and multivariate data analysis
All strains were separately inoculated in Petri dishes containing PDA medium, and incubated for 7 days at 25 °C in order to ensure that all of them were of the same age.Then, one pellet (diameter 6 mm) of the strain was transferred to vials (40 mL) containing 10 mL of PDB, and immediately sealed with septa of silicone and threaded caps.After incubation for 14 days at 25 °C under static conditions, the vials were placed in a bath of ethylene glycol at 60 °C, and the VOCs were extracted for 30 min by HS-SPME using a PDMS/DVB fiber placed above (ca. 1 cm) the surface of the fungal culture.After this period, the fiber was removed and inserted in the GC-MS at 250 °C for 4 min for the VOC desorption.A vial containing only PDB (no fungus) was used as the control.For the optimization of the HS-SPME conditions, a 2 2 trial planning with two quantitative variables (temperature and extraction time) was carried out at two levels (50 and 60 °C; 10 and 30 min, all in duplicate), with a central point (60 °C and 20 min, in triplicate).Eleven experiments were performed, and the statistical analysis was done by using the program Quality Tools: Statistics in Quality Science. 23nalyses of GC-MS were performed on a QP-2010 gas chromatograph coupled to a mass spectrometer from Shimadzu ® (Tokyo, Japan), with a capillary DB-5 column (25 m × 0.32 mm × 0.5 µm) from J & W Scientific ® (Folsom, USA).The analysis conditions were as follows: injector temperature 250 °C; GC oven temperature 35 °C for 2 min, from 35 to 195 °C (20 °C min −1 ), from 195 to 220 °C (10 °C min −1 ), and from 220 to 280 °C (20 °C min −1 ); mode of injection 1:5 split; volumetric flow rate of the carrier gas (Helium) 0.59 mL min −1 ; detector temperature 250 °C.Mass spectra were obtained by electron impact (70 eV) in the range of m/z 18 to 400 (intervals 0.5 s).The VOCs were identified by the obtained mass spectra with those from mass spectral libraries (NIST 05, NIST 27, Wiley 229 and Adams 24 ), and by the calculated linear retention indexes with literature data. 19,24,25The GC-MS of the VOCs were subjected to PCA and HCA analyses by using the free software R Project from R Foundation for Statistical Computing ® (Vienna, Austria). 23The analyzed data correspond to the averages of injections in triplicate.The matrix arrangement was made up of ten lines (number of fungi analyzed) and thirty-four columns (compounds identified).The matrix of Pearson correlation was used to perform the PCA.To the HCA, Euclidean distance was used as the coefficient of dissimilarity, and the grouping was done by the method of association average (Ward).The option of automatic truncation was chosen to define the conglomerates and to obtain the dendrogram.

Results and Discussion
The HS-SPME was used on the extraction of the VOCs produced by ten species of endophytic fungi from the Botryosphaeriaceae family.The extraction of volatile compounds was optimized by performing the experimental design 2 2 , considering the temperature (50 and 60 °C) and extraction time (10 and 30 min) as variables.The highest number of compounds was detected when the experiment was performed at 60 °C and 30 min.
As already reported by Valente and Augustos, 26 both time and temperature affect the method of extraction by HS-SPME, influencing the kinetics of mass transfer of the volatile compounds between the different phases of the system, and the thermodynamics, which describes the partition equilibrium of the VOCs.In order to verify the significance of these two variables on the extraction of VOCs, all data were analyzed by the application of the Pareto graph method (Figure 1).It was observed that the variable time (B) has greater significance than the variable temperature (A), and that lower significance was observed when the two variables were together.
The VOCs from all strains were extracted under the optimized conditions and analyzed by GC-MS.Thirty-four volatile compounds were identified as being produced by the fungal strains (Table 1), and not observed in the control experiments.Most of the compounds are sesquiterpenes (14 non-oxygenated and 10 oxygenated), and only two linear ketones and eight alcohols were identified.Strain 683 (Neofusicoccum ribis) produced the greatest number of identified compounds (26), while strain 20 (Botryosphaeria mamane) had the lowest number of identified VOCs (6 compounds).
Only three monocyclic sesquiterpenes (β-elemene, α-bisabolol and germacrene D) were found.The nonoxygenated sesquiterpenes belong mostly to the cadinane group while the oxygenated ones possess the guaiane and cedrane skeleton.No monoterpenes were identified in the investigated strains.Seven primary alcohols with different side chain lengths, an aromatic alcohol (phenylethyl alcohol), and two ketones (undecan-2-one and hexadecane-3-one) were also identified.PCA (Figure 2) followed by HCA (Figure 3) were applied to the identified VOCs in order to investigate the chemical variability patterns.All compounds were considered, and the averages of triplicates were treated as independent variables in a similarity matrix.According to Figure 1, which shows a plot of the scores of the 34 variables from the ten samples, it was possible to differentiate all species.In this case, two main components explain 47.89% of the variance, while five main components explain 77.71% of the total.P. stromaticum (strain 474) presented a greater separation from the other strains due to the presence of α-bisabolol as the main component in high concentrations (88.83%).The dendrogram of HCA (Figure 3) corroborates the differentiation of the ten investigated species.With respect to N. parvum (strain 600) and N. ribis (strain 683), both PCA and HCA clearly differentiated them as two distinct species, and corroborated the use of VOCs analysis as an auxiliary tool for the identification of these species.
The high content (88.83%) of the sesquiterpene, α-bisabolol, produced exclusively by P. stromaticum (strain 477), suggests its potential as a biomarker.This compound is naturally occurring in plants and was first isolated from Matricaria chamomilla (Asteraceae). 27-Bisabolol has been used in cosmetic formulations and has important biological activities, such as anti-inflammatory, anti-irritant, antibacterial, antispasmodic, anti-allergic, drug permeation and vermifuge. 27,28Although α-selinene has been found in most of the investigated strains (except strains 277 and 477), the high content (33.70%) of this sesquiterpene in B. mamane (strain 20) may suggest it as a biomarker for this species.As already mentioned, N. parvum (strain 600) and N. ribis (strain 683) are closely related, and it is not trivial to differentiate them even by molecular analysis.In this study, both species produced α-cecrene epoxide and guaiol acetate, but in different concentrations.α-Cecrene epoxide (38.99%) was produced in high percentage by strain 600, while guaiol acetate (59.89%) was the major constituent of strain 683.Thus, these sesquiterpenes may be considered biomarkers for differentiating these two species.

Conclusions
The HS-SPME in association with GC-MS was successfully used for the study of the VOCs produced by ten filamentous fungi species from the Botryosphaeriaceae family.The VOCs profiles of the investigated strains proved to be adequate to differentiate them by multivariate data analysis (PCA and HCA).In addition, N. parvum and N. ribis, which were previously considered very similar species, were also differentiated through their volatile chemical profiles.The high content of α-bisabolol in P. stromaticum suggested this compound as a potential chemotaxonomic marker for this species.As far as we know, this is the first report on the VOCs profile of the ten investigated Botryosphaeriaceae species.

Figure 1 .
Figure 1.Pareto chart for the effects of temperature and time on the VOCs extraction.

Figure 2 .
Figure 2. PCA scores plot of 34 variables from samples of HS-SPME-GC-MS of ten fungi from the Botryosphaeriaceae family.

Figure 3 .
Figure 3. Dendrogram of HCA of the HS-SPME-GC-MS of ten fungi from the Botryosphaeriaceae family.

Table 1 .
VOCs produced by endophytic fungi from the Botryosphaeriaceae family