Untargeted and Targeted LC-MS/MS Based Metabolomics Study on In Vitro Culture of Phaeoacremonium Species

Grapevine (Vitis vinifera L.) can be affected by many different biotic agents, including tracheomycotic fungi such as Phaeomoniella chlamydospora and Phaeoacremonium minimum, which are the main causal agent of Esca and Petri diseases. Both fungi produce phytotoxic naphthalenone polyketides, namely scytalone and isosclerone, that are related to symptom development. The main objective of this study was to investigate the secondary metabolites produced by three Phaeoacremonium species and to assess their phytotoxicity by in vitro bioassay. To this aim, untargeted and targeted LC-MS/MS-based metabolomics were performed. High resolution mass spectrometer UHPLC-Orbitrap was used for the untargeted profiling and dereplication of secondary metabolites. A sensitive multi reaction monitoring (MRM) method for the absolute quantification of scytalone and isosclerone was developed on a UPLC-QTrap. Different isolates of P. italicum, P. alvesii and P. rubrigenum were grown in vitro and the culture filtrates and organic extracts were assayed for phytotoxicity. The toxic effects varied within and among fungal isolates. Isosclerone and scytalone were dereplicated by matching retention times and HRMS and MS/MS data with pure standards. The amount of scytalone and isosclerone differed within and among fungal species. To our best knowledge, this is the first study that applies an approach of LC-MS/MS-based metabolomics to investigate differences in the metabolic composition of organic extracts of Phaeoacremonium species culture filtrates.


Introduction
Vitis vinifera is one of the most economically important crops worldwide, with approximately 71% of the world's grape production being used for wine production. A variety of fungal diseases threatens viticultural regions all over the world, compromising the yield and quality of wine [1,2]. Out of them, grapevine trunk diseases (GTDs), caused by one or several xylem-inhabiting fungi, produce a progressive decline in vines, consisting of a loss in productivity and eventually death of the vines [3]. Over the past few decades, they have been extensively studied [4]. However, the relationship between pathogenic fungi involved in GTDs and abiotic agents, the expression of symptoms, and the lack of effective management strategies, requires further investigation [5,6]. The most common GTDs include Esca and Petri diseases, Botryosphaeria, Diaporthe and Eutypa diebacks,

Phaeoacremonium Species and Culture Conditions
Nine representative isolates belonging to three different Phaeoacremonium species were selected. In details, four isolates of P. italicum (CBS 137763, extype culture; Pm50M; Pm59; Pm45) were retrieved from the collection of the Department of Science, Food Natural Resources and Engineering (DAFNE), University of Foggia (Foggia, Italy). Three isolates of P. alvesii (CBS 113590; CBS 729.97; CBS 408.78) and two isolates of P. rubrigenum (CBS 498.94 extype culture; CBS 112046) were retrieved from Westerdijk Fungal Biodiversity Institute (CBS, Utrecht, The Netherlands). Taxonomic and morphological identity of all Phaeoacremonium species were confirmed and discussed according to Laidani et al. (2021) [43].

Extraction Procedure
At harvest, the mycelial mat was removed from each flask by filtration under vacuum with sterile filter paper. Extractions were carried out at two different pHs: (1) unmodified pH (UpH) of culture filtrates (pH ranged from 4.4 to 8.8 among the species tested); (2) pH 2, by acidification with 1 M formic acid (Sigma-Aldrich, Milan, Italy). For each fungal isolate, two samples of 40 mL were taken from the culture filtrates and their pH modified. Then each culture filtrate was extracted three times using ethyl acetate (3 × 40 mL) (Sigma-Aldrich, Milan, Italy). Organic extracts corresponding to the same pH value were then combined, dried over Na 2 SO 4 (Sigma-Aldrich, Milan, Italy), filtered, and evaporated under reduced pressure. The organic extracts obtained from each fungal isolate and pH value, and their corresponding aqueous phases, were tested for phytotoxicity as explained in the following section.

Phytotoxic Bioassays
The culture filtrates and the organic extracts (UpH and pH 2) were assayed on non-host cotyledons of Cucumis sativus L. Before the inoculation, the cotyledons were disinfected with EtOH 70% for 45 sec, subsequently rinsed three times with sterile distilled water and finally dried in sterile paper. For the culture filtrates, two droplets (20 µL) were directly spotted on each cotyledon without dilution. Droplets (20 µL) of sterile water and of Czapek broth were used as negative controls. Meanwhile, the organic extracts were tested at the concentration of 3 and 1.5 mg/mL. The organic extracts were dissolved in MeOH and diluted with distilled H 2 O up to the assay concentrations (the final content of MeOH was 4%). Droplets of the test solutions (10 µL) were applied on the axial side of cotyledons that had been needle punctured 3-4 times. Droplets (10 µL) of Czapek broth and MeOH in distilled H 2 O (4%) were applied on cotyledons as negative controls. Both experimental bioassays were repeated twice. The cotyledons were then kept in a moist chamber to prevent the drying of the droplets, observed daily and scored for symptoms after five and eight days. The presence of symptoms obtained by organic extract bioassays was evaluated by observation using a scale of 0-5, where 0 = no symptoms observed; 1 = 1-20%; 2 = 21-40%; 3 = 41-60%; 4 = 61-80%; and 5 = 81-100% of cotyledon surface showing necrotic symptoms. The data recorded were used to determine the overall toxicity severities (TS) according to Equation (1):

Untargeted and Targeted LC-MS/MS Analysis
Chromatographic separations were achieved by two LC-MS platforms based on either a Q-Exactive Hybrid Quadrupole-Orbitrap (Thermo Fisher, Waltham, MA, USA) coupled with an Infinity 1290 UHPLC System (Agilent Technologies, Santa Clara, CA, USA), or a hybrid triple quadrupole/linear ion trap tandem mass spectrometer (QTRAP 4500, AB Sciex, Framingham, MA, USA) coupled to an Eksigent Ekspert ultraLC UPLC. The same chromatographic conditions were applied to both platforms by using a C18 column (Eclipes Plus, 3.5 um, 4.6 × 100 mm) from Agilent (Santa Clara, CA, USA) at 28 • C and an elution program consisting of a linear gradient from 10% to 80% of 0.1% (v/v) formic acid in ACN containing 0.1% (v/v) formic acid in 6 min. A post run equilibration step of 4 min was included prior to each analysis. The injection duty cycle was 10 min considering the column equilibration time. Flow rate was 0.4 mL/min. The injection volume was 5 µL and the autosampler was maintained at 10 • C.
For the untargeted analysis, the Q-Exactive-based instrument was equipped with a Heated Electrospray (HESI) source with the following setting: spray voltage positive polarity 3.2 kV, negative polarity 3.0 kV, capillary temperature 320 • C, S-lens RF level 55, auxiliary gas temperature 350 • C, sheath gas flow rate 60, and auxiliary gas flow rate 35. Full MS scans were acquired over the mass range 150-500 with a mass resolution of 70,000. The target value (AGC) was 1 × 10 6 and the maximum allowed accumulation time (IT) was 100 ms. For the data dependent MS/MS (ddMS2) analyses a Top10 method was used. The ten most intense peaks were selected for fragmentation with a stepped normalized energy of 25-28 and 20-30 in positive and negative ionization mode, respectively. AGC was 2 × 10 5 with IT 75 ms and 17,500 mass resolution. Injection volume was of 5 µL.
Targeted MS analysis was performed on hybrid triple quadrupole/linear ion trap tandem mass spectrometer QTRAP 4500. Q1 resolution was adjusted to 0.7 ± 0.1 amu for multi reaction monitoring (MRM), referred to as unit resolution. Q3 was also set to unit resolution in MRM mode. MS analysis was carried out in positive ionization mode using an ion spray voltage of 4500 V. The nebulizer and the curtain gas flows were set at 30 psi using nitrogen. The Turbo V ion source was operated at 400 • C with the auxiliary gas flow (nitrogen) set at 50 psi. Two suitable multi reaction monitoring (MRM) transitions were selected for the compounds scytalone (1) and isosclerone (2), while one transition was selected for the I.S. 2-hydorxy-1,4-naphtoquinone (3). The compound dependent parameters for the three compounds were optimized using the manual optimization protocol in tuning mode. The Q1 mass, the Q3 transition and the best parameters are reported in Table 1.
Validation study was obtained analysing calibration curves, limit of detection (LOD), limit of quantification (LOQ), within-day and between-day imprecision and inaccuracy. Calibration curves were obtained reporting the Area ratio (Compound Area/I.S. Area) against the compound concentration. I.S. was spiked in every standard solution in a concentration of 5 µg/mL. Two different curves were built for scytalone (1) and isosclerone (2), respectively (Table S1). Each solution was injected in triplicate. LOD and LOQ were calculated as a magnitude of, respectively, 3 and 10 times the standard deviation of noise to the lower point of standard level [45] for both compounds (Table S1). Two quality control samples (QC) at two different concentration levels were used for assessing the within-day (Table S2) and between-day variation ( Table S3). The within-day imprecision (Coefficient of Variability (CV) %) and inaccuracy (%) were calculated analysing, in the same analytical run, each level of QC samples 6 times (Table S2). The between-day imprecision (CV %) and inaccuracy (%) were calculated analysing each level of QC samples once a day, for 5 days (Table S3). tion of noise to the lower point of standard level [45] for both compounds (Table S1). Two quality control samples (QC) at two different concentration levels were used for assessing the within-day (Table S2) and between-day variation ( Table S3). The within-day imprecision (Coefficient of Variability (CV) %) and inaccuracy (%) were calculated analysing, in the same analytical run, each level of QC samples 6 times (Table S2). The between-day imprecision (CV %) and inaccuracy (%) were calculated analysing each level of QC samples once a day, for 5 days (Table S3).

Data Analysis
Instrument control, data acquisition of MS spectra acquired by UHPLC-Orbitrap, were performed using the associated Excalibur software version 4.0. Instrument control, data acquisition, and processing of MS spectra acquired by UPLC-QTrap were performed using the associated Analyst and MultiQuant Software version 1.6 and 3.0.2, respectively. Raw UPLC-Orbitrap data were converted to mzML file format using the open-source software ProteoWizard [46]. Data processing was carried out using the software MZMine 2 [47]. The data processing workflow included: (1) mass detection; (2) ADAP chromatogram builder [48]; (3) smoothing; (4) deconvolution; (5) deisotoping; (6) alignment (RANSAC Aligner); (7) gap filling. More details about the parameters are available in Table S4. Statistical analysis that included Principal Component Analysis (PCA), Partial Least Squares-Discriminant Analysis (PLS-DA) and Hierarchical Clustering were conducted using MetaboAnalyst© 5.0 [49,50]. A normalization to sample median, square root data transformation and autoscaling was applied. Metlin [51], NpAtlas [52] and m/z mine [53]  tion of noise to the lower point of standard level [45] for both compounds (Table S1). Two quality control samples (QC) at two different concentration levels were used for assessing the within-day (Table S2) and between-day variation ( Table S3). The within-day imprecision (Coefficient of Variability (CV) %) and inaccuracy (%) were calculated analysing, in the same analytical run, each level of QC samples 6 times (Table S2). The between-day imprecision (CV %) and inaccuracy (%) were calculated analysing each level of QC samples once a day, for 5 days (Table S3).

Data Analysis
Instrument control, data acquisition of MS spectra acquired by UHPLC-Orbitrap, were performed using the associated Excalibur software version 4.0. Instrument control, data acquisition, and processing of MS spectra acquired by UPLC-QTrap were performed using the associated Analyst and MultiQuant Software version 1.6 and 3.0.2, respectively. Raw UPLC-Orbitrap data were converted to mzML file format using the open-source software ProteoWizard [46]. Data processing was carried out using the software MZMine 2 [47]. The data processing workflow included: (1) mass detection; (2) ADAP chromatogram builder [48]; (3) smoothing; (4) deconvolution; (5) deisotoping; (6) alignment (RANSAC Aligner); (7) gap filling. More details about the parameters are available in Table S4. Statistical analysis that included Principal Component Analysis (PCA), Partial Least Squares-Discriminant Analysis (PLS-DA) and Hierarchical Clustering were conducted using MetaboAnalyst© 5.0 [49,50]. A normalization to sample median, square root data transformation and autoscaling was applied. Metlin [51], NpAtlas [52] and m/z mine [53]  tion of noise to the lower point of standard level [45] for both compounds (Table S1). Two quality control samples (QC) at two different concentration levels were used for assessing the within-day (Table S2) and between-day variation ( Table S3). The within-day imprecision (Coefficient of Variability (CV) %) and inaccuracy (%) were calculated analysing, in the same analytical run, each level of QC samples 6 times (Table S2). The between-day imprecision (CV %) and inaccuracy (%) were calculated analysing each level of QC samples once a day, for 5 days (Table S3).

Extraction of Phaeoacremonium Culture Filtrates
Four isolates of Phaeoacremonium italicum, three isolates of P. alvesii and two isolates of P. rubrigenum were tested for their ability to produce phytotoxins in vitro, as reported in the material and methods section. The results for all these samples are reported in Table 2. The highest yield in organic extract was obtained in the range between 7.13 and 28.15 mg per 40 mL filtrate when the culture filtrates were extracted at acid pH (Table 2).

Phytotoxic Bioassays
Symptoms on cotyledons treated with culture filtrates appeared 5 days after the inoculation. Necrotic spots, browning of cotyledon veins, and irregular discoloured areas were observed (Figure 1). In particular, among the P. italicum, the isolate Pm59 showed both necrotic spots and browning of veins symptoms, while isolates CBS 137763, Pm50M and Pm45 showed just browning of veins or necrotic spots. Whereas, P. alvesii isolates CBS 113590, CBS 408.78 and CBS 729.97, showed necrotic spots and browning of veins. P. rubrigenum isolates CBS 498.94 and CBS 112046 showed light necrotic spots and irregular discoloured areas. No symptoms were observed on cotyledons assayed with Czapek broth and sterile distilled water ( Figure 1).
The symptoms that occurred on cotyledons assayed with organic extracts at 3 mg/mL started to appear at the fifth day after inoculation, and the toxicity severity increased during the following 3 days. All the Phaeoacremonium species extracts caused necrotic spots on cotyledons regardless of the concentration and the pH of extraction, while the appearance of necrotic areas, discoloured areas and chlorotic ring depended on species, pH, and concentration ( Table 3). The symptoms caused by the extracts of P. italicum Pm50M and Pm45, P. alvesii CBS 729.97 and P. rubrigenum CBS 498.94 varied depending on extraction pH values (UpH and pH 2). In detail, the extracts at unmodified pH (UpH) of P. italicum Pm50M and P. alvesii CBS 729.97 caused necrotic spots surrounding the puncture, irregular discoloured areas and marginal necrotic areas, whereas only necrotic spots were observed on cotyledons treated with pH 2 extracts. The extracts at pH 2 of P. italicum isolate Pm45 and P. rubrigenum isolate CBS 498.94 showed necrotic spots surrounding the puncture, irregular discoloured areas and marginal necrosis, while necrotic spots and irregular discoloured areas were observed on cotyledons treated with UpH extracts. The remaining 10 extracts gave the same symptoms regardless of the pH of extraction (Table 3; Figure 1). Similar symptoms were observed for the culture filtrates treated at 1.5 mg/mL, even though they were less pronounced. No symptoms were observed on cotyledons assayed with Czapek broth and MeOH (4%).
The toxicity severity (TS) scores, calculated as reported in the materials and methods section (Equation (1) filtrates treated at 1.5 mg/mL, even though they were less pronounced. No symptoms were observed on cotyledons assayed with Czapek broth and MeOH (4%). The toxicity severity (TS) scores, calculated as reported in the materials and methods section (Equation (1) (2) For the dereplication procedure, standard solutions of scytalone (1) and isosclerone (2) were analysed by UHPLC-Orbitrap in polarity switching mode. The best ionization condition for both compounds was found in the negative mode, thus the HRMS data, the MS/MS fragments, and retention time (RT), were annotated (Table 1). After filtration, 20 extracts, including 18 extracts of the fungal isolates and 2 extracts of the culture medium, were analysed in duplicate under the same conditions of the standards. Figure 2 shows examples of the Total Ions Chromatograms (TICs) of the extracts at unmodified pHs and under acid conditions (pH 2).    The raw spectra were then visualised by Excalibur software that allowed rapid dereplication of scytalone (1) The presence of compound 1 was confirmed in the extracts of both P. alvesii isolates CBS 729.97 and CBS 408.78, P. italicum isolates CBS 137763, Pm45, Pm59 and P. rubrigenum isolate CBS 112046. Furthermore, compound 1 was mainly detected in the UpH extracts (Table 4). Isosclerone (2) was detected in the extracts of P. italicum isolates Pm45 and CBS 137763, P. alvesii isolate CBS 408.78 and P. rubrigenum isolate CBS 498.94. Compound 2 was detected only in the in the UpH extracts. ---a UpH = extracted at same pH of the culture filtrate; b pH 2 = extracted after acidification; c <LOD = the amount is below the limit of detection d <LOQ = the amount is below the limit of quantification. (1) and Isosclerone (2) The scheduled MRM method (sMRM) consists in a short analytical cycle of 10 min ( Figure S1). Two different calibration curves were developed for scytalone (1) and isosclerone (2). The LODs and the r values are reported in Table S1. Before proceeding with the analysis of the extracts, two levels of QC samples were prepared to evaluate imprecision and inaccuracy of the method. The within-day imprecision (CV%) and inaccuracy (%) were calculated by analysing, in the same analytical run, each level of QC samples six times. The imprecision varied from 2.2% to 4.2% for scytalone (1) and 1.9% to 3.1% for isosclerone (2). The inaccuracy (%) ranged from −5.7% to 5.5% for scytalone (1), and from −0.8% to −0.8% for isosclerone (2) ( Table S2). The between-day imprecision (CV%) and inaccuracy (%) were calculated analysing each level of QC samples once a day, for 5 days. The imprecision is estimated from (CV%) 5.6% to 7.6% for scytalone (1), from 2% to 6.3% for isosclerone (2); the inaccuracy was estimated from (%) 1.4% to 1.75% for scytalone (1) and from 1.8% to 2.3% for isosclerone (2) (Table S3). After the validation the 20 extracts were analysed in triplicates and the quantitative data analysed by Multiquant software. The concentration of scytalone (1) was higher in all the UpH extracts (Table 3) and was >LOQ in the following isolates: P. italicum Pm45 and CBS 137763, P. alvesii isolates CBS 729.97 and CBS 408.78 and P. rubrigenum isolates CBS 112046. P. italicum Pm45 UpH extract showed the higher concentration of 1 (11.90 µg/mL ± 1.05), while the lowest concentration (1.00 µg/mL ± 0.12) was found in P. rubrigenum isolate CBS 112046 (Table 4). Isosclerone (2) was quantifiable only in P. italicum isolates Pm45 and CBS 137763. The highest concentration (86.67µg/mL ± 5.08) of 2 was found in P. italicum CBS 137763. Finally, only P. italicum isolates CBS 137763 and Pm45 produced both compounds in quantifiable amounts.

Statistical Analysis Using MetaboAnalyst
MzMine 2 was used for the processing of UHPLC-Orbitrap spectra as described in the materials and method section. All the parameters used for the processing are reported in Table S4. The obtained peak lists were analysed by MetaboAnalyst 5.0 using the statistical analysis module.
PCA was used to observe an overview of variance between the extracts at different pH and secondary metabolite compositions. The PCA score plot showed the Phaeoacremonium species crude extracts clustered according to the extraction condition (Figure 3a). However, the PCA showed no differences among or between the species. This result was also confirmed by the hierarchical clustering, shown as a dendrogram in Figure S2. Thus, even though it was possible to cluster the organic extracts depending on the extracting conditions, it was not possible to separate the species according to their metabolic profile. The box plots of scytalone (1) Figure 3b, show a high concentration of these two compounds in the UpH extracts. These data further confirm the results arose from the quantitative analysis (Table 4).

Selection of Phaeoacremonium Isolates for Further Investigation
Several unknown compounds can be spotted in the untargeted analysis but the comparison of their MS and MS 2 data with those reported on web-based databases was unsuccessful. In order to start a first characterization of these compounds, one isolate of each species was selected for additional in vitro growth. All the fungal organic extracts showed phytotoxicity in the assay condition, thus the selection of the fungal isolates was carried out, also comparing the results of multivariate statistical analysis and the metabolic profile obtained by UHPLC-Orbitrap.
Thirteen out of fifteen of the top features in the VIP score (Figure 3d) were most Partial Least Squares-Discriminant Analysis (PLS-DA) was carried out to investigate the significant differences among metabolites extracted at different pH, and to further explore possible differences among the species. The PLS-DA score plot (Figure 3c) confirms the PCA results: the crude extracts were again grouped according to the extraction condition. The statistical model showed a good predictivity having Q2 and R2 values, obtained using Pareto scaling at five components of 1 and 0.9, respectively. The variable importance of projection (VIP) score scatter plot indicating the top 15 features, responsible for the clustering of the groups, was shown in Figure 3d

Selection of Phaeoacremonium Isolates for Further Investigation
Several unknown compounds can be spotted in the untargeted analysis but the comparison of their MS and MS 2 data with those reported on web-based databases was unsuccessful. In order to start a first characterization of these compounds, one isolate of each species was selected for additional in vitro growth. All the fungal organic extracts showed phytotoxicity in the assay condition, thus the selection of the fungal isolates was carried out, also comparing the results of multivariate statistical analysis and the metabolic profile obtained by UHPLC-Orbitrap.
Thirteen out of fifteen of the top features in the VIP score ( Figure 3d) were most abundant in the UpH extracts. Therefore, the metabolic profiles of organic extracts of fungal isolates at UpH, grouped by species (Figure 4a-c), were compared to select the most promising samples. According to the biological assay, isolate Pm45 had the highest TS score (1.63) among the P. italicum. As shown in Figure 4a, this isolate had also the richest metabolic profile among this species. Interestingly the peak at 3.54 min showed the same high-resolution mass of scytalone (1) but different MS 2 pattern, which is suggestive of a possible structural isomerism. The structure elucidation of this compound is currently under investigation. Within the P. alvesii species, the isolate CBS 113,590 had the highest TS score (2.13). The isolate CBS 729.97 showed the second TS score (1.50) and produced the highest amount of scytalone (1). This isolate was also characterized by the richest metabolic profiles, with a very diagnostic peak at 10.76 and m/z 167.0127 that is currently under study for structure elucidation (Figure 4b). Within the P. rubrigenum species the isolate CBS 498.94 had the highest TS score (2.00) and the richest metabolic profile. In this sample, the most abundant peak occurred at 5.89 min and showed a molecular mass at m/z 198.0173. As for the other unknown compounds reported above, this product is also currently under investigation (Figure 4c). Figure 4d shows a comparison of the chromatograms of the selected isolates, some of the peaks were shared between the isolates, even though with different abundance. Other peaks seem to be specific for a selected isolate (Figure 4d).
profile. In this sample, the most abundant peak occurred at 5.89 min and showed a lecular mass at m/z 198.0173. As for the other unknown compounds reported above, product is also currently under investigation (Figure 4c). Figure 4d shows a compar of the chromatograms of the selected isolates, some of the peaks were shared betw the isolates, even though with different abundance. Other peaks seem to be specific f selected isolate (Figure 4d).

Discussion
To the best of our knowledge, this is the first study reporting the dereplication quantification of scytalone (1) and isosclerone (2) produced in vitro by Phaeoacremon

Discussion
To the best of our knowledge, this is the first study reporting the dereplication and quantification of scytalone (1) and isosclerone (2) produced in vitro by Phaeoacremonium species applying untargeted and targeted LC-MS based metabolomics approaches. Prior to the present analysis, no data were available in literature on the secondary metabolites produced by P. italicum, P. alvesii and P. rubrigenum. Moreover, the untargeted metabolic profiles of their organic extracts were investigated for the presence of other secondary metabolites. Two different pH conditions, unmodified pH and pH 2 were used for the extraction of secondary metabolites. The number of organic extracts obtained results higher for pH 2 according to previous reported data in literature for other pathogenic fungi involved in grapevine trunk diseases [31,32]. The pH is a parameter that can affect the extraction of secondary metabolites, due to their diverse nature and physicochemical properties. Consequently, an extraction condition suitable for one chemical class may be unsuitable for another.
The choice of the best extraction condition plays a dominant role in the comprehensiveness and the representativeness of the metabolite profile obtained, and sometimes could be challenging [54]. This aspect is fundamental especially when multi-omics studies want to be applied. Different extraction procedures optimized for different biomolecules are needed, and a flexible extraction method providing robust and reliable recovery of the molecular components is advisable [55].
Both culture filtrates and organic extracts used in bioassays produced different phytotoxicity degrees, although the symptom kinds were similar. In particular, necrotic spots, browning of cotyledon veins, and irregular discoloured areas were more intense in the culture filtrates rather than in the organic extracts. The higher phytotoxicity of the culture filtrates could be related to the presence of high molecular weight phytotoxins in the culture filtrates, which could have a synergist effect intensifying the observed symptoms [56], and/or to the presence of active fungal propagules passed through out the filter paper. Indeed, toxic exopolysaccharides and polypeptides have been isolated from Phaeoacremonium minimum, one of the main causal agents of Esca complex culture filtrate [57,58]. Thus, future studies could be conducted to investigate high molecular weight phytotoxins produced by Phaeoacremonium species to understand their role in pathogenicity.
Phaeoacremonium minimium is also known to produce two phytotoxic naphthalenone pentaketides: scytalone (1) and isosclerone (2) [29]. They were also identified in liquid culture filtrates of Phaeomoniella chlamydospora also involved in Esca complex [30]. For this reason, 1 and 2 were dereplicated in the organic extracts of the three selected Phaeoacremonium species by matching retention times, and HRMS and MS/MS data, with pure standard samples.
Dereplication strategies are needed to screen the crude extracts for the presence of known compounds before isolation efforts are initiated. In fact, re-isolation of known compounds is a crucial problem, wasting time and resources, in the discovery of new biologically active compounds from natural sources [42]. To tackle this problem, modern dereplication strategies have been developed. They use high throughput screening techniques, such as tandem mass spectrometry, coupled with bioinformatics tools and databases, both in house built and web based [59,60]. The progress of metabolomics and molecular networking also helped in the development dereplication strategies [61].
Scytalone (1)  To gain further insights into the production of 1 and 2 a quantitative targeted LC-MS method has been developed on UPLC-QTrap and then validated. QTrap together with triple quadrupole (QqQ) are the work horses in targeted approaches. Both instruments rely on MRM (Multi Reaction Monitoring) as their most sensitive and reliable method in the quantitation of metabolites [62]. As previous attempts to quantify scytalone (1) and isosclerone (2) from culture filtrates of P. minimum by LC-MS failed [63], this study reports for the first time a valid targeted MS method for their quantification in in vitro culture of Phaeoacremonium species. In details, scytalone (1) was detectable in both pH extracts, although the highest amount was found in the unmodified pH, while isosclerone (2) was measurable only in the UpH extracts. This difference could be explained considering the chemical features of scytalone (1). Compound 1, having two phenolic groups, is more acid then isosclerone, thus could be extracted also at pH 2, although in smaller quantities than UpH. These results highlighted that the pH is a key parameter for the extraction of 1 and 2, and, with high probability, also for close related metabolites produced by this genus [28]. P. italicum isolate Pm45 produced the highest amount of scytalone (1). A quantifiable amount of 1 was also detected in P. italicum CBS 137763, P. alvesii isolates CBS 729.97 and CBS 408.78, and P. rubrigenum isolate CBS 112046. Isosclerone (2) was measurable only in P. italicum isolates Pm45 and CBS 137763. Furthermore, combining these quantitative data with the biological assay results, no correlation between the amount of 1 and 2 and toxicity severity scores was found, because they were independent from the amount of scytalone and isosclerone found in the extracts. Hence, also in this case, synergistic or antagonistic effects [56] could be present in the extracts and should be studied.
These results point out the importance of structural identification of the components of the organic extracts. The pre-treatment of the untargeted LC-MS data was carried out by MzMine 2.0 [47], while the statistical analysis was carried out by MetaboAnalyst [50]. The principal component analysis (PCA) clustered the crude extracts into two groups according to the extraction pH values. These data further confirm the importance of pH value as a key parameter for the extraction process. On the contrary, no differences were highlighted in the extracts of distinct species. Moreover, the PCA confirmed the highest amount of scytalone (1) and isosclerone (2) in the UpH. Partial Least Squares-Discriminant Analysis (PLS-DA) further confirmed the previous results. Organic extracts clustered according to the pH values, but regardless of the species. Although the multivariate analysis did not show differences between the selected species, this workflow could be easily applied to compare the metabolic profile of other pathogenic fungi involved in the Esca complex or, in general, other fungi associated with GTDs. The main goal of these studies could be the identification of potential species-specific biomarkers.
The differences between the unmodified UpHs and pH 2 were mainly due to unknown metabolites; the attempt to dereplicate features having a VIP score >1.55 using online database failed. The annotation of small molecules remains a major pitfall in untargeted LC-MS-based metabolomics. The association of a detected individual signal to its corresponding metabolite identities is often performed by querying detected signals against one or several experimental web-based databases. Outstanding progress has been made in the last decade in growing the number of metabolites in databases; however, these databases are far from being comprehensive. Researchers involved in in spectroscopic characterization play an important role in filling this gap, and they should be prone to uploading the collected spectra in metabolomics databases [64].
Only by elucidating unknown metabolites is it possible to biologically interpret complex systems, gaining insight into the chemical-ecology of pathogenic fungi. Assignment of the chemical structure to fungal unknown secondary metabolites could be beneficial for investigations that deal with fungal pathogenic behavior [65], host-pathogen interaction [66][67][68] or, more in general, that use multi-omics approaches [69]. Consequently, three isolates (Pm45; CBS 729.97; CBS 498.94), one for each of Phaeoacremonium species, were selected according to the toxicity severity score, multivariate statistical analysis and untargeted metabolic profile, to be grown in vitro in larger amounts (4 to 8 L), to achieve the isolation and chemical (by spectroscopic method) and biological characterization of the unknown secondary metabolites. The spectroscopic data collected will be uploaded on the most used metabolomic database.

Conclusions
Investigations on the bioactive natural products produced by fungi have a significant impact in many scientific research fields, from plant pathology to organic chemistry. Using novel tools such as metabolomics can help to better understand and interpret complex biological phenomena, such as host-pathogen interaction. Scytalone (1) and isosclerone (2), and a few other naphthalenone pentaketides have been isolated from Phaeoacremonium species to date [28]. However, the untargeted metabolic profiling disclosed several unknown compounds that could be related to other secondary metabolites that could play a role in the life cycle of this species. Moreover, species of Phaeoacremonium have been reported to be responsible for infections in humans [11], thus investigating the role of secondary metabolites in pathogenesis could also be noteworthy in medicine applications. Indeed, as described by the One Health concept [70], human health is linked to the health of the ecosystems of which we are a part.