2.2. Screening of Epigenetic Modifiers Effects on Fungal Endophytes
All strains were grown in the same production medium (YES) in the absence (control) and presence (100 μM) of seven small-molecule epigenetic modifiers that were added: (i) only during the production stage (− +), or (ii) during both, seed and production, fermentation stages (+ +). According to previous reports [10
], all epigenetic modifiers were tested at 100 μM as the standard concentration in the study. After 14 days of fermentation, ensuring the generation of enough biomass in the control conditions, submerged culture differences in morphology and SMs profiles were analyzed visually and by UHPLC-UV at 210 nm. Changes in production titers (as changes in 210 nm UV areas) and presence of new peaks were observed among the different treatments with respect to their corresponding controls. Increasing changes in production profiles were classified into three categories as p-pp-ppp
(as increasing changes: ×2, ×4 and ×8 respectively, in the area of the peaks under the uHPLC UV-210 nm trace compared to its corresponding control areas). Changes in the chemical diversity induced were also categorized based on the number of new UHPLC-UV 210 nm peaks detected as d
: 1–3 peaks; dd
: 3–5 peaks and ddd
> 5 peaks (Table 2
). In addition, any changes in the fermentation morphology (pigmentation, final biomass or conidia/hyphae conversion rate), were also referred as m
in the table.
Three of the 13 fungal strains screened by using the 14 epigenetic modifiers fermentation conditions showed: clear significative changes in their final growth morphologies, changes in their production titers and/or generation of new peaks when cultured in the presence of the epigenetic modifiers. Among them, the strain that showed more changes in these three components was the strain Dothiora
sp. CF-285353. This strain was therefore selected for a deeper and more detailed metabolomic study on the effects of the three elicitors that induced most morphological and SMs profile changes when compared to its corresponding control fermentation condition (Figure 2
2.3. Standardization of Cultivation Conditions of Dothiora sp.
The study on the effect of the epigenetic modifiers has shown that the two DNMT inhibitors, 5-azacitydine (1) and hydralazine hydrochloride (2), and the sirtuin activator quercetin (4), induced more changes in the SMs profiles of the strain CF-285353 when added both, to the seed and production stages (+ +), than when added only during the production stage (− +). To verify that these changes were significative and reproducible, extensive studies were designed and carried out with submerged cultures per triplicate, in the presence and absence of these three elicitors, both in the seed and/or the production stages.
The cultivated strain CF-285353 presented different morphologies after addition of some of the epigenetic modifiers (Figure 2
). To ensure the reproducibility of the experimental conditions, we characterized the morphology of the strain both in the inoculum and in the different fermentation steps. We considered, as well, the risk of introducing mutations and strain degeneration after each transfer and cell division by the liquid-liquid sub-culturing required to obtain enough inoculum for the study [16
]. This effect has been previously observed in studies comparing jasmonic acid production by a fresh and a sub-cultured producing strain, which concluded that liquid-liquid sub-culturing led to a significative downfall of jasmonic acid production by Lasiodiplodia theobromae
Therefore, three experimental designs where tested for minimizing these inoculum sub-culturing effects: (i) a method involving the use of one single stage seeds from agar plugs, similar to the original process performed during the previous screening of the epigenetic modifier additives (Figure 3
A); (ii) a second method where a pre-inoculum enrichment step was added for generating a large amount of inoculum from a unique production batch (Figure 3
B); and (iii) a third one involving a pre-inoculum, preserved as frozen stocks, from which agar plates of the endophyte with a controlled differentiation morphology were prepared, and were later used to inoculate seeds for each production batch (Figure 3
Finally, and similarly to the screening procedure, each inoculum generated was then seeded in the production medium (YES), in absence (control) and presence (100 μM) of the three best small-molecule epigenetic modifiers selected: (i) added only during production or; (ii) added during all fermentation steps, seed and production.
2.3.1. Effects of Epigenetic Modifiers on Dothiora sp. Submerged Culturing
To assess the effect of the epigenetic modifiers, all fungal cultures were harvested after 14 days of incubation, and were characterized according to their morphology (color, biomass and mycelial-to-yeast conversion phase; Figure 4
), extracted with organic solvent and analyzed by UHPLC-MS.
Regarding the morphology, clear differences were observed between the three control fermentation batches for each sub-culturing methods tested even without any epigenetic modifier. The fermentation triplicates performed using the initial screening approach (Figure 3
A) presented, consistently, black thick broth with a high density of conidia (Figure 4
A), whereas fermentation controls for the other two inoculating methods (Figure 3
B,C) resulted in a yellowish broth with plenty of hyphae and few conidia (Figure 4
B,C). A slightly higher amount of hyphae was present when using the frozen inocula (Figure 3
C), even when harvested after at 7 days, half of the time used for the other two methods (14 days). No other morphology differences could be related to the presence or absence of any of the epigenetic modifiers, thus all three methods and the two growth morphologies were decided to be included in further metabolomic differential analyses.
2.3.2. Metabolomic Evaluation of the Changes in the SM Profile of Dothiora sp. by Epigenetic Modifiers
HPLC-UV traces can be used for evaluating the effect of epigenetic modifiers on fungal strains and have proved to be suitable for a fast detection of the conditions that induce strong changes in their SM profiles. In our case, extended studies required deeper analytical methods to quantify and identify the SMs that could be induced by the addition of the modifiers. We decided to use mass spectrometry analyses, as metabolomics is currently getting stronger foundation on LC/MS data. In recent studies, Volcano-plots have been applied to these data with success in describing and obtaining conclusions on the biosynthesis of fellutamides by A. nidulans
, hence supporting the use of HDAC inhibitors and said techniques for the discovery of cryptic secondary metabolites [12
Volcano-plots constitute a scatter-plots representation that can describe very visually how two different experimental conditions may affect a large set of components. Statistically, these plots determine if significative differences exist between averages of two populations of the same component treated with the two conditions of interest, depicting these results for, in our case, the secondary metabolites of a given extract (or the mass ions that can be detected by LCMS and can be inferred as components).
sp. fermentations (Figure 3
and Figure 4
), we observed molecular species whose production increased or decreased significantly with respect to the corresponding controls for each condition (see Figure 5
for a detailed number of the ions statistically different in production for each modifier and fermentation method). In general, the addition of small-molecule elicitors, both during the inocula and the production fermentations, resulted in a significant increase of the diversity and amount of secondary metabolites generated for the three epigenetic modifiers of the study. For the two seeding methods that have included a pre-inoculum step, 5-azacytidine (1
) was the epigenetic modifier that induced more changes when added both during the seed and production fermentations, and the number of significant unique molecular species detected by HPLC-MS at least doubled when compared to its addition only during the production fermentation.
In most of the cases we observed a continuous dispersion of the p-values for the significance of the differential components for confidences above 95% and 99%. This indicated a general and continuous epigenetic modifiers effect on the SMs profiles, without highlighting any outlier group of ions that could indicate a unique strongly induced secondary metabolite pathway. Thus, results confirmed the general non-specificity of the effects induced by the modifiers on the culture broths SMs profiles. In fact, only hydralazine was observed to induce clear groups of outlier mass ions in the scatter-plots.
The conditions showing less production changes on the SMs abundances compared to their controls were the fermentations that were prepared with the pre- and cryo-inocula method (Figure 5
C). In contrast, the higher dispersions were obtained with the initial production method (Figure 5
A). Although fermentations were harvested with similar biomasses, the method that included pre- and cryo-inocula steps presented the lowest dispersion of their volcano-plots (Figure 5
C), suggesting an important influence of the methodology used.
2.3.3. Identification of Molecules Produced in the Fermentations with Modifiers
As previously commented, hydralazine (2
) was the epigenetic modifier that induced most of the differential mass ion populations for CF-285353. In fact a total number of 23 and 99, 6 and 11, and 10 and 17 ions were identified respectively as statistically differential, with a 99% of significance, for treatments with hydralazine in (− +) and (+ +) for methods A, B and C (see Figure 5
for details). Among them, we could identify some molecules with 8 to 32 fold higher production rates in the presence of this modifier, and some others that were only produced in its presence.
In an initial attempt to identify several of the natural products induced by this modifier, we compared every ion that presented a statistical significance above 99% to our internal de-replication UV-HPLC-HRMS databases. Database matching was performed by using an in house developed application where the UV signal, retention time, mass signal and moleculat formula of the selected ions are compared to the UV-HPLC-HRMS data of known metabolites stored in our proprietary database of 845 microbial natural products, including commercial compounds and molecules obtained from internal purification campaigns. Among the metabolites with increased production, we could identify the presence of curvicollide A/B (m/z 432; C26H40O5) and fusidic acid (m/z 516; C31H48O6), and more tentativelly (matching of their molecular formulas with the commercial database Dictionary of Natural Products, DNP): pyrophen (m/z 309; C18H17NO5), cyclo-isoleucyl-leucyl-isoleucyl-leucyl (m/z 452; C24H44N4O4), melledonal C (m/z 497; C24H29ClO8) and rhizoxin S (m/z 613; C34H47NO9), previously described to be produced by a endosymbiotic bacteria in Rhizopus spp. On the contrary, when compared to the control condition, the production of several secondary metabolites was repressed in the presence of hydralazine (2). These melecules were identified tentatively, by comparison to our databases and DNP, as monascuspyrone (m/z 340; C19H32O5), pleurotin (m/z 354; C21H22O5), roseotoxin B (m/z 591; C30H49N5O7) and 12-hydroxy-8,10-octadecadienoic acid (m/z 653; C18H32O3). It is important to mention that all these molecules were only identified tentatively and future chemically directed purifications and HRMS/NMR confirmations are needed for a definitive confirmation.
Most differential ions detected with the volcano-plots methodology were not found in our chemical de-replication databases and could not be identified. These ions, according to their highest to lowest statistical significance p-value, included, for the production method A) (− +): m/z 243–244, 185, 374–375, (+ +): m/z 266–267–268, 243–244–245, 185–186–187; Production method B) (− − +): m/z 185–186, 157, 243, 247 and 225, (− + +): m/z 185–186, 171, 243, 247 and 227; and production method C) (− − +): m/z 185–186, 243 and 431, (− + +): 185–186, 243, 225, 171 (continuous ions, that are listed here together with hyphens, may belong to high intensity low resolution LC/MS signals detected in the raw data as contiguous m/z ions). In general, some of these most outlier ion masses were induced independently of the production method applied, being four of them the most statistically significative: m/z 171, 185–186, 225–226–227 and 243–244–245. These four sets of ions were studied then by UHPLC/HRMS-MS and their molecular formulae identified as C9H6N4 (9), C10H8N4 (10), C13H14N4 (11) and C14H16N2O2 (12). Interestingly the first three outlier molecules presented a molecular formula closely related to that of hydralazine (C8H8N4, 2), and could not be found also in public nor commercial natural products databases. Further scale up and purification studies were setup for their identification.
Medium volume (600 mL) fermentations of the strain in the presence and absence of hydralazine (2
) were extracted and fractionated for the isolation and identification of these differential molecules. HRMS/NMR results indicated that m
185–186, and m
227 ions corresponded to biotransformation products (9
) of hydralazine (2
) ocurring in the fermentation broth. On the other hand, the molecular formula C14
, deduced from the ions m
243–244–245, that could be associated to 19 possible molecules in the natural products databases, was finally purified and identified as the diketopiperazine natural product cyclo(phenylalanyl-prolyl) (12
) by HRMS and NMR data (Figure 6
and Supplementary Materials
2.3.4. Dothiora sp. Biomarkers Related to Growth Morphology
The differential growth morphologies observed in the strain CF-285353 and the availability of the metabolomic tools implemented for the evaluation of the epigenetic modifiers effects, also prompted us to perform the comparison of the profiles generated by the strain in the different morphologies, with the aim of identifing possible morphology biomarkers that could help us to monitor and fine tune future Dothiora sp. fermentations.
A representative population of eight CF-285353 extracts that presented the yellow hyphal morphology were compared with a similar population of extracts that presented the black conidial morphology, by using the volcano-plots approach (Figure 7
). Statistically, and with a 99% of confidence, 713 mass ions were present in the conidial growths with significative higher titers than in the corresponding hyphal growths. On the other hand, 237 mass ions were present significatively in higher amounts in the hypha than in the conidia growth. Additionally, clear outlier populations of ions were observed for each condition, highlighting several SM pathways differentially expressed according to the fermentation morphology presented by the fungus.
The outlier ion masses with more significative abundance in the hyphal morphology were m
268, 269 and 226, whereas significative abundant ion masses in the conidial morphology were m
222, 465, 486 and 503. Both sets of components were studied, identified and their structures confirmed by UHPLC/HRMS-MS to correspond respectively to the primary metabolite adenosine (13
) and the natural product mycosporin: glutamicol-5′-O
) (Figure 7
Previous studies on the morphology, growth kinetics, and main chemical components of solid or liquid cultures of Tolypocladium
fungi isolated from wild Cordyceps sinensis
, showed that the in vitro
hyphae mycelium of Tolypocladium
presented much higher contents of adenosine (1116.8 μg vs.
264.6 μg) than when the fungus was directly isolated from C. sinensis
]. Similarly to what was observed for Tolypocladium
, adenosine could represent a potential biomarker of the hyphae growth morphology state as observed in our strain. Previous work by Pirttila et al.
] suggests that the adenosine secretion of endophytes microorganisms may play an important role in the morphological development of the host plant and therefore, it could be playing a role in the secondary metabolism of the endophytes microorganism.
In the case of mycosporins, extremophilic microcolonial fungi have been described to constitutively synthesize considerable amounts of mycosporins, also known to be involved in morphogenesis and sporulation [20
]. Many reports regarding different yeast species indicate the ability of these fungi to synthesize mycosporin-like amino acids, which have been also proposed to reflect phylogenetic relationships among species, suggesting their utility in yeast systematics. In addition, mycosporins have also been described in the extracellular matrix and in the outer cell wall layers of microcolonial fungi, in which they mediate a wide range of intracellular reactions. They are present in the mucilage that surrounds conidia of some fungi, confirming its direct presence in cultures with this morphology, with no evidences of their presence inside the conidia [23
]. In some fungi, mycosporins are also described to protect conidia from solar radiation during atmospheric dispersal and prevent untimely germination [24
] prolonging their survival. Therefore, the modulated production of these molecules by yeasts represents an interesting subject for further research due to its ecological, taxonomical and biotechnological implications [25
]. For example, cosmetics applications as UV protectors and activators of cell proliferation, with potential therapeutic properties, could be interesting for other commercial developments [26
Both detected molecules, adenosine (13) and mycosporin glutamicol-5′-O-β-d-glucopyranoside (14), confirmed in the literature as related to these morphology growth stages of some fungi, also corroborate the volcano-plots methology as a suitable approach for a succesful identification of metabolites produced differentially between two fermentation conditions, allowing a fast and robust approach for the identification of fermentation biomarkers.