OSMAC approach leads to new fusarielin metabolites from Fusarium tricinctum

Abstract

Using the OSMAC (One Strain MAny Compounds) approach, the fungal endophyte Fusarium tricinctum was cultivated on fruit and vegetable juice-supplemented solid rice media. This led to an up to 80-fold increase in the accumulation of the new natural product fusarielin J (1), as well as to the induction of two new natural products fusarielin K (2) and fusarielin L (3) and the known derivatives fusarielins A (4) and B (5). Compounds 2–5 were not detected when the fungus was grown on rice media lacking either fruit or vegetable juice. The highest increase in the accumulation of compound 1 was observed in the presence of apple and carrot juice, whereas the stimulating effect was weaker for banana juice. Compound 1 exhibited cytotoxicity against the human ovarian cancer cell line A2780, with an IC₅₀ value of 12.5 μM.

Introduction

The fungal genus Fusarium is most prominently known for its phytopathogenic effects on various staple food crops, causing depletory ‘Fusarium wilt’ (Panama Disease) in bananas¹ and ‘Fusarium head blight’ in cereals, such as wheat, oat, barley and maize.² Fusarium species (for example, F. graminearum and F. tricinctum) produce mycotoxins that are potentially harmful to humans, whereas some of them can also be used as drugs or drug candidates. These mycotoxins include beauvericins, trichothecenes and most importantly enniatins^{3, 4} that were medicinally used (fusafungine) up to the beginning of this year as local antibiotics in nasopharyngitis treatment.⁵ Enniatins are currently being investigated as potential anticancer agents,^{6, 7} and so are semisynthetic derivatives of the cyclic depsipeptide sansalvamide A, another Fusarium metabolite with strong in vitro cytotoxicity against pancreatic, colon and breast cancer cell lines.^{8, 9, 10, 11} In addition, Fusarium strains have transiently produced several antineoplastic agents formerly known from plants such as paclitaxel, camptothecin and L-asparaginase,^{12, 13, 14} as well as podophyllotoxin,¹⁵ the latter being the lead compound that led to the clinically used semisynthetic drugs etoposide and teniposide.

These examples demonstrate how auspicious secondary metabolites from Fusarium species are and that ongoing natural product research on this genus promises to be a highly rewarding strategy for the discovery of new bioactive compounds.¹⁶

One of the approaches for drug discovery from fungi involves the application of the OSMAC (One Strain MAny Compounds) method that attempts to induce silent biogenetic clusters and hence may lead to accumulation of compounds not accumulated during more conventional fermentations. OSMAC terms the modulating effect that altered culture conditions (that is, media composition, temperature, osmolarity, pH) may have on the secondary metabolite production of microorganisms.^{17, 18} Examples for such culture variations for filamentous fungi include the use of different liquid or solid media, as applied in a previous study with F. tricinctum fermented on solid beans and liquid Wickerham medium versus cultivation on solid rice medium, leading to an enhanced production of enniatins on beans.^{19, 20} Another example is the mimicry of extreme habitats by cultivating at colder temperatures or using highly saline media (for example, growing Gymnascella dankaliensis on either NaCl- or NaBr-enriched rice medium) that led to the induction of chlorinated or brominated metabolites.^{21, 22, 23}

With Fusarium species such as F. graminearum and F. tricinctum, addition of CaBr₂ to the cultivation medium led to an induced production of the bioactive compound fusarielin A.^{24, 25} Fusarielin-type polyketides are a therapeutically promising class of Fusarium metabolites: with 17 known derivatives identified so far (fusarielin A–I, ICM0301 A–H),^{26, 27, 28, 29, 30} fusarielins display antifungal,²⁸ mild antibacterial^{24, 31} and anti-angiogenic activities,^{27, 32} in addition to weak cytotoxicity against HeLaS3 and NCI-H69 cells.²⁶ Interestingly, previous cultivation experiments have shown enhanced fusarielin yields by cultivating F. tricinctum on complex carbon sources.²⁵ Therefore, having an endophytic F. tricinctum strain at hand that was previously isolated from the rhizome of Aristolochia paucinervis, we decided to implement the OSMAC concept by cultivating it on solid rice medium supplemented with three types of commercially available fruit and vegetable juices, namely apple, banana and carrot juice, as these naturally contain a high percentage of various carbohydrates. Herein, cultivating F. tricinctum on rice containing juice had a starkly enhancing and inducing effect on the production of fusarielins: the new fusarielin J (1) was enhanced 80-fold compared with control cultures lacking juice and the production of the new fusarielins K (2) and L (3) and the known fusarielins A (4) and B (5) was induced. A total of 14 natural products were identified based on their NMR and MS spectra and all fusarielins (Figure 1) were investigated for their cytotoxicity against A2780 human ovarian cells.

Figure 1

Structures of compounds 1–5.

Results and Discussion

F. tricinctum was cultivated on solid rice medium with or without fruit or vegetable juice in two flasks per medium variation. In each case the fermentation was stopped by addition of ethyl acetate after 14 days and all experiments were carried out in triplicate, resulting in 6 cultivation flasks per medium.

The ethyl acetate extracts of all treatments presented typical metabolic profiles in HPLC analysis (Figure 2): the known compounds (−)-citreoisocoumarin (6), (−)-citreoisocoumarinol (7), lateropyrone (8), enniatin B (9), enniatin B1 (10), enniatin A1 (11), enniatin A (12) and fusaristatin A (13) were readily identified by LC-MS analysis and by comparison with authentic standards that had either previously been isolated from the same fungal strain,³³ or in the case of phenyl acetic acid (14) by comparison with a commercially available standard. The amounts of 6–13 per flask irrespective of the presence or absence of juice were comparable to those of previous experiments (Table 1).³³ Compound 14 was isolated from extracts of juice-supplemented cultures; however, it was not detectable in the HPLC chromatograms. Interestingly, the juice-based cultivations enhanced the production of the new natural product fusarielin J (1) up to 80-fold compared with fungal cultures lacking juice (Table 1). Furthermore, the cultivation of F. tricinctum with juice induced the production of two new natural products fusarielin K (2) and fusarielin L (3), and of two known natural products fusarielins A (4) and B (5) that were not detected in fungal controls. None of the investigated compounds including the new natural products 1–3 were detected when only the juices used for the described experiments were analyzed.

Figure 2

HPLC chromatograms (min vs mAU) overlay of EtOAc extracts: Fusarium tricinctum on solid rice medium prepared with water (a), banana (b), apple (c) and carrot juice (d); compound 14 was not detectable. A full color version of this figure is available at the Journal of Antibiotics journal online.

Table 1 Compound yields per flask (n=6)

Compound 1 was isolated as a white, amorphous solid and exhibited a UV absorption maximum at 241 nm. The high-resolution electrospray ionization mass spectrometry (HRESIMS) of 1 displayed a prominent ion peak at m/z 561.3182 [M+Na]⁺ indicating the molecular formula C₃₃H₄₆O₆, consistent with 11 degrees of unsaturation. In the ¹H NMR spectrum of 1, 43 protons were detected, including two pairs of overlapping aromatic protons at δ_H 7.26 (brd, 7.4, H-4′/8′) and 7.34 (brt, 7.4, H-5′/7′), one aromatic proton at δ_H 7.30 (brt, 7.4, H-6′), four olefinic protons at δ_H 6.18 (dd, J=15.0, 10.8, H-6), 5.85 (dd, J=10.8, 1.2, H-5), 5.29 (qd, J=6.7, 1.5, H-19) and 5.02 (dd, J=15.0, 10.2, H-7), five downfield shifted aliphatic protons at δ_H 4.70 (brq, J=2.2, H-11), 3.78 (d, J=9.0, H-3), 3.55 (dd, J=10.6, 5.8, H_α-1), 3.70 (dd, J=10.6, 5.0, H_β-1) and 2.74 (brs, H-15), suggesting their direct vicinity to an oxygen atom, three methylene groups at δ_H 1.53 (t, J=14.0, H_α-13) and 1.65 (m, H_β-13), 1.27 (m, H_α-10) and 1.40 (m, H_β-10), and 3.62 (s, H₂-2′), five methine protons at δ_H 2.55 (d, J=5.2, H-17), 2.18 (m, H-8), 1.91 (ddd, J=14.0, 11.1, 3.0, H-14), 1.80 (m, H-2) and 1.55 (m, H-9), and six methyl groups at δ_H 1.69 (d, J=1.2, H₃-22), 1.67 (d, J=6.7, H₃-20), 1.61 (brs, H₃-25), 1.21 (s, H₃-24), 1.04 (s, H₃-23) and 0.78 (d, J=6.9, H₃-21). The COSY spectrum revealed four spin systems that consisted of H-21/H-2/H₃-21/H-3 (A), H₃-22/H-5/H-6/H-7/H-8/H-17/H-9/H₂-10/H-11/H₂-13/H-14 (B), H-19/H₃-20/H₃-25 (C) and H-4′/H-5′/H-6′/H-7′/H-8′ (D) (Figure 3a). After assigning all protons to their attaching carbon atoms via analysis of the HSQC spectrum (Table 2), the aforementioned spin systems were connected through interpretation of the heteronuclear multiple bond correlation (HMBC) spectrum as follows: the first spin system (A) was connected to the second (B) by the correlation from H-3 to the quaternary carbon C-4 (δ_C 137.4) and CH₃-22 (δ_C 11.5), and by the J³ correlation from H-5 to the oxygenated C-3 (δ_C 81.9). The third spin system (C) was elucidated as a but-2-ene moiety through a J² correlation from H₃-25 to the quaternary carbon C-18 (δ_C 134.7) and J³ correlations to C-19 (δ_C 126.1). Moreover, the HMBC correlation from H₃-25 to C-17 and from H-8 to C-18 connected this moiety (C) to the decalin system at position C-17. By further HMBC evaluation and comparison with the known fusarielin B (5), which was also isolated in this study, the basic framework of 1 was found to be identical to that of 5.²⁶

Figure 3

(a) Key COSY correlation (bold lines) and heteronuclear multiple bond correlation (HMBC; arrows) of 1; (b) key HMBC (plain arrows) and NOE (dashed arrows) correlations of the 2,2,5-trimethyl-1,3-dioxane moiety of 2. A full color version of this figure is available at the Journal of Antibiotics journal online.

Table 2 NMR spectroscopic data of 1 [CD₃OD] (¹H NMR in 600 MHz, ¹³C in 150 MHz and δ in p.p.m.)

The remaining signals (spin system D and H₂-2′) were attributed to a phenyl acetate unit, as supported by the HMBC correlations from the aromatic protons H-4′, H-5′, H-7′ and H-8′ to C-3′ (δ_C 135.7) and the methylene carbon C-2′ (δ_C 42.7) and from H-6′ to C-3′. Similarly, the methylene protons H₂-2′ correlated with C-3′ as well as with the aromatic carbons C-4′/8′ and C-5′/C7′ (δ_C 129.4) and the carbonyl carbon C-1′ (δ_C 171.8). The connection of the phenyl acetate function to the fusarielin B (5) framework was established through the HMBC correlation from H-11 to C-1′ (Figure 3a), and was corroborated by the marked downfield shift (Δ 1.43 p.p.m.) of this proton, as a result of deshielding effects from the resulting ester function. The relative configuration of 1 was deduced by the J-values extracted from the ¹H NMR spectrum (Table 2) and by interpretation of the 2D NOESY spectrum (Figure 3a) to be identical to that of fusarielin B (5). Thus, compound 1 was identified as a new natural product and was named fusarielin J.

Compound 2 was isolated as a white, amorphous solid and showed a similar UV–visible spectrum as 1. The molecular formula C₃₆H₅₀O₆ was deduced by the prominent ion peak at m/z 596.3944 [M+NH₄]⁺ in the HRESIMS, suggesting an additional structural functionality of 40 amu compared with 1. The ¹H NMR spectrum of 2 was very similar to that of 1, indicating that both compounds shared the same fusarielin core structure. Additional signals were detected at δ_H 1.34 (s, H₃-26) and 1.48 (s, H₃-27), assigned to two methyl groups. In the HMBC spectrum of 2, the correlations of both CH₃-26 and CH₃-27 with each other as well as with the remaining ketal quaternary carbon C-1′′ (δ_C 99.3) revealed the presence of an additional acetonide moiety. The HMBC correlation from H2-1 to C-1′′ confirmed the position of the acetonide group through ether linkages at C-1 and C-3 (Figure 3b), rationalizing the additional degree of unsaturation. The relative configuration of the fusarielin core structure of 2 was determined to be identical to that of 1 on the basis of the respective coupling constants and NOESY correlations (Table 3). In addition, the six-membered dioxane ring adopted a chair conformation (Figure 3b), as suggested by the NOE correlations of H₃-27 with the axial protons H-1α and H-3, indicating their co-facial orientation. This was further corroborated by the ¹³C chemical shifts of CH₃-26 and CH₃-27 (δ_C 29.9 and 19.4, respectively) that are characteristic of a syn 1,3-diol acetonide moiety. The acetonide function is known from other fungal natural products; for example, ganodermacetal from the basidiomycete Ganoderma amboinense,³⁴ or cochliomycins A and B from the ascomycete Cochliobolus lunatus.³⁵ The isolation of 2 was performed in the absence of acetone, and the fact that no acetonide was formed, even after incubating 1 mg of 2 in 1 ml acetone for 24 h strongly suggested that 2 is a natural product and not an artifact arising from 1 during the isolation procedure. Therefore, compound 2 was identified as a new natural product and was given the name fusarielin K.

Table 3 NMR spectroscopic data of 2 [CD₃OD] (¹H NMR in 600 MHz, ¹³C in 150 MHz and δ in p.p.m.)

Compound 3 was isolated as a colorless, amorphous solid and displayed a similar UV spectrum as 1 and 2. The molecular formula of 3 was identified as C₂₈H₄₂O₄ based on its prominent ion peaks detected at m/z 465.2978 [M+Na]⁺, revealing a 40 amu increase in the molecular weight compared with fusarielin A (4). The ¹H and ¹³C NMR data of 3 were similar to those of 4, apart from the presence of two additional methyl signals at δ_H 1.33 (s, H₃-26) and 1.46 (s, H₃-27), thus suggesting the presence of an additional acetonide moiety as found in 2. This assumption was further corroborated by careful analysis of the HMBC spectrum of 3, in which the cross peaks from H_β-1 (δ_H 3.72), H₃-26, and H₃-27 to C-1′ (δ_C 99.4) were discerned. The relative configuration of 3 was established on the basis of NOE correlations, coupling constants and ¹³C NMR data (Table 4), by analogy with 1 and 2. Therefore, 3 was identified as the 1,3-diol acetonide derivative of 4 and was given the name fusarielin L.

Table 4 NMR spectroscopic data of 3 [CD₃OD] (¹H NMR in 600 MHz, ¹³C in 150 MHz and δ in p.p.m.)

The known compounds fusarielins A (4) and B (5) were identified by comparison of their spectroscopic data with those in the literature.²⁶ In addition, the absolute configurations of 4 and 5 (Figure 1) had previously been deduced on the basis of CD spectroscopy and chemical derivatization experiments.²⁶ Therefore, based on the similar NMR data and specific optical rotation values of 1, 2 and 3 ([α]²⁰_D -76, -53 and -33, respectively) compared with those of 4 and 5 ([α]²⁰_D -132 and -38, respectively) as well as on their close biogenetic relationship, it is assumed that these compounds (1–5) share the same absolute configuration (Figure 1).

In a previous study on F. tricinctum, the influence of different carbon and nitrogen sources, as well as variation of pH, cultivation time and temperature had been analyzed.²⁵ Interestingly, the use of complex carbon sources showed the strongest enhancement of fusarielin A (4) production.²⁵ As juices are naturally high in carbohydrates (6–14 g per 100 ml), our experiments clearly underpin these findings. Moreover, our juice experiments led to an additional enhancement and/or induction of more and new fusarielin derivatives (1–3). However, the carbohydrate content (banana>apple>carrot) of each juice was not coherent with the yield of fusarielins per flask (apple>carrot>banana) (Table 1). Therefore, other juice components/factors may play an additional role in fusarielin induction and enhancement. In order to pinpoint which physical characteristic(s) of the juices (also known as OSMAC effector) comports linear to the induction and enhancement of fusarielins, we compared several properties of the juices and media, including pH, osmolarity and juice contents, but no interrelations were found (Table 5). This is not surprising as previous OSMAC studies (for example, using tap water versus distilled water) revealed that even slight changes in the cultivation medium, such as differences in metal ion concentrations, trigger the production of new bioactive secondary metabolites in endophytic fungi.³⁶

Table 5 Nutritional and physical characterization of liquids and rice medium used for cultivation

All fusarielins (1–5) were submitted to a cytotoxicity assay against the human ovarian cancer cell line A2780. Compound 1 exhibited moderate cytotoxicity with an IC₅₀ value of 12.5 μM against the A2780 cells. All other fusarielins (2–5) showed very weak activity (>36 μM) in the same assay. For the known fusarielins A (4) and B (5), this was in accordance with previously published data.²⁶ Interestingly, the acetonide function in fusarielin K (2) and L (3) led to a three- and twofold decreased activity compared with the precursors fusarielins J (1) and A (4) (Table 6). The phenyl acetic acid function seems to play an important role in the cytotoxic activity of fusarielin J (1) and K (2). Compound 1 is fourfold more cytotoxic than phenyl acetic acid lacking fusarielin A (4) and B (5), and 2 is about twofold more cytotoxic than fusarielin L (3).

Table 6 IC₅₀ values of compounds 1–5 analyzed in an MTT assay against A2780 cells

To the best of our knowledge, this is the first report using fruit or vegetable juices for an OSMAC approach, and from the fusarielin derivatives identified in this study so far only fusarielin A (4) had been isolated from cultures of F. tricinctum before.^{24, 25, 37} In addition, we were able to induce the production of the known fusarielins A (4) and B (5), and of two new fusarielins K (2) and L (3), along with the 80-fold enhancement of the new fusarielin J (1). In conclusion, employing juice as an addition to the fermentation media of fungi proves to be a promising OSMAC approach for the discovery of new bioactive secondary metabolites in the ongoing quest of natural product research to provide model structures for the development of potential future drugs.

Experimental Procedure

General experimental procedures

For a measurement of OR of chiral compounds, the latter were dissolved in optically pure solvents (Uvasol, Merck Chemicals, Darmstadt, Germany) and left to equilibrate for 30 min. The specific OR was then measured thrice and averaged for each sample, using a PerkinElmer 241MC polarimeter (Perkin-Elmer, Waltham, MA, USA). All NMR data (1D and 2D) were measured in deuterated methanol on Bruker Avance III 300, Bruker Avance DRX 500 or Bruker Avance III 600 MHz NMR spectrometers at 25 °C (Bruker BioSpin, Rheinstetten, Germany), and calibrated with the deuterated methanol peak (δ_H 3.31 p.p.m. and δ_C 49.0 p.p.m.). The molecular weight of each compound was determined using an HP110 Agilent Finnigan LCQDecaXP Thermoquest ESI mass spectrometer (Thermo Fisher Scientific, Bremen, Germany), and the molecular formula was assigned with the help of HRESIMS spectra measured on a UHR-TOF maxis 4G mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). All fractions and pure compounds were monitored via analytical HPLC DIONEX p580 UVD340S PDA DAD3000RS (Dionex Softron, Munich, Germany).³⁸ Compounds were purified via semipreparative HPLC Merck Hitachi HPLC, I-7100 pump and I-7400 detector (Merck KGaA, Darmstadt, Germany)³⁸ or preparative TLC polygramSIL G/UV₂₅₄, pre-coated (Macherey-Nagel, Düren, Germany), using HPLC grade eluents (Merck Chemicals) in each case. Initial and subsequent fractionation of crude extracts was performed using column chromatography with stationary phase either silica gel 60M (Macherey-Nagel)³⁸ or Sephadex LH-20 25–100 μm bead size (Sigma-Aldrich, Munich, Germany). TLC plates were used to monitor the collected fractions by spotting them at 254 and 366 nm. The osmolarity of the juices was measured with an osmomat3000 osmometer (Gonotec, Berlin, Germany). The pH values of the rice media were determined using pH-indicator strips (Merck Chemicals).

Microbial material

F. tricinctum had previously been isolated as an endophyte from rhizomes of Aristolochia paucinervis collected in Beni-Mellal (Morocco).³³ Fresh mycelia of F. tricinctum were grown on agar plates³⁹ and used for all cultivation experiments.

Culture conditions

Aqueous solid rice culture control medium was prepared following our standard operation protocol.³⁹ For the OSMAC study, the demineralized water in the mentioned protocol was replaced with banana juice (Pfanner, St Peter, Germany), apple juice (Naturkind, Kaiser’s Tengelmann, Mülheim an der Ruhr, Germany) and carrot juice (Naturkind) by volume measurement (110 ml per 100 g rice, autoclavation at 121 °C for 20 min). Each medium was prepared in duplicate, resulting in eight Erlenmeyer flasks in total per experiment that was performed three times. Per flask, half an agar plate mycelium was added under sterile conditions, using four fully overgrown plates per experiment in total. The cultures were left to grow at room temperature (21 °C) and in the dark. After 14 days, the mycelia of F. tricinctum had completely covered and grown through each medium, and thus were harvested at this time. This experiment was repeated twice.

Extraction and isolation

Each flask was infused with 600 ml EtOAc, the solid medium was chopped up inside the flask and after covering each flask up, they were extracted on a horizontal shaker (140 r.p.m.) for 8 h. Subsequently the flasks were left in a static state overnight before being filtered and dried over a rotary evaporator to generate the EtOAc crude extract per flask. The mean dry weight yield of the aqueous cultures amounted to ca. 180 mg per flask, of the banana juice cultures to ca. 275 mg per flask, of the apple juice cultures to ca. 307 mg per flask and of the carrot juice cultures to ca. 353 mg per flask. For initial characterization, each extract was dissolved in 50 ml MeOH (HPLC grade, VWR, Radnor, PA, USA) and injected into the above-mentioned analytical HPLC with 20 μl per extract. The extracts of the apple and carrot juice cultures were combined and then submitted to vacuum LC separation. As these yielded a sufficient quantity, extracts from banana juice were not included. The fraction obtained with the eluent n-hexane/EtOAc (1:4) amounted to 290 mg and was further separated over a Sephadex column with methanol. Five of the so generated subfractions were further worked on: fraction 3 (96 mg) was purified via preparative TLC (methanol/DCM, 1:4) yielding compounds 1 (18.8 mg), 4 (5.2 mg), 2 (3.3 mg), 5 (7.2 mg) and 3 (1.4 mg).

Compounds 7–14 were identified from the crude extracts via LC-MS analysis with the corresponding known compounds previously isolated in our group.³³ An experimental calibration curve (concentration vs area units; R² >0.9) was produced for each purified compound by HPLC measurements of several concentrations, so that the yields per flasks for each compound could be calculated from the resulting curve function.

Fusarielin J (1)

White, amorphous, hygroscopic solid; [α]²⁰_D -76° (c 0.14, MeOH); UV (MeOH) λ_max (log ɛ) nm 243 (14.7); ¹H, ¹³C and 2D NMR data Table 2; HRESIMS m/z 561.3182 [M+Na]⁺ (calcd for C₃₃H₄₆O₆, 561.3184).

Fusarielin K (2)

White, amorphous, hygroscopic solid; [α]²⁰_D -53° (c 0.14, MeOH); UV (MeOH) λ_max (log ɛ) nm 243 (14.1); ¹H, ¹³C and 2D NMR data Table 3; HRESIMS m/z 596.3944 [M+NH₄]⁺ (calcd for C₃₆H₅₄NO₆, 596.3946).

Fusarielin L (3)

Colorless solid; [α]²⁰_D -33° (c 0.11, MeOH); UV (MeOH) λ_max (log ɛ) nm 242 (15.0); ¹H, ¹³C and 2D NMR data Table 4; HRESIMS m/z 465.2978 [M+Na]⁺ (calcd for C₂₈H₄₂O₄Na, 465.2975).

Biological assays

All fusarielins (1–5) isolated in this study were investigated in an MTT assay against the cisplatin-sensitive A2780 human ovarian cancer cell line, following previous protocols.^{40, 41}