Metabolomics and Dereplication Study of the Endophytic Fungus Aspergillus chevalieri in Search of Bioactive Natural Compounds

Objectives: This study aimed to point the significant rule of metabolomics tools to assess the chemistry of the bioactive metabolites produced by endophytic fungus Aspergillus chevalieri isolated from Lagerostromia tomentosa C. presl. The anticancer of crude extracts, fractions and pure compounds and antimicrobial of pure compounds were investigated as part of this study. Methods: An endophytic fungus (Aspergillus chevalieri) was isolated from the tissues of the stem of Lagerostromia tomentosa C. Presl and identified through molecular biological procedure by DNA isolation, PCR, DNA sequencing and through searching the Gene Bank. Metabolomics profiling and dereplication studies were employed to choose the optimum growth medium and conditions that yield the most significant metabolites. The crude extract of the 30-days rice culture of Aspergillus chevalieri was subjected to bioactivity and metabolomics guided isolation approach. The structure of the isolated compounds was determined on the basis of 1D, 2D NMR and mass spectrometry (HR-ESIMS) analysis. Results: four fractions were further purified to produce five pure compounds, which are Ergosterol (1), Ergosterol peroxide (2), Campesterol (3), Flavoglaucin (4) and 3-Omethyl caffeic acid (5). Multivariate data analysis highlighted the most significant metabolites contributed to the bioactivity. The pure compounds were tested for the anticancer and antimicrobial activity, compound (1) exhibited significant antitrypanosomal activity, while compounds (2, 3, 4 and 5) effectively inhibited the growth of Escherichia coli, Staphylococcus aureus and Candida albicans. Conclusion: A combination of metabolomicand bioassay-guided approaches gives an access to a shorter and faster route to highlight the active metabolites, which are highly correlated to the bioactivity during the first stage of fractionation.


INTRODUCTION
Lagerstroemia is an important member of family Lytheraceae, which contains thirty one genera.Many species of genus Lagerstroemia are known to have an important medicinal uses as anti-oxidant, antibacterial, antiviral, anti-inflamatory, antinociceptive, anti-diarrhoeal, cytotoxic, xanthine oxidase inhibition and anti-obesity, also the leaf extract shows anti-diabetic activity due to presence of ellagitannins and corosolic acid 1 .
Aspergillus is one of the most popular and common endophytic fungi found to be associated with natural sources either marine or terrestrial habitat, it belongs to Ascomycetes fungi 2,3 .Endophytes are a highly diverse group of fungi capable of living symbiotically inside plant tissue without causing apparent symptoms of diseases 4 .Endophytes might be involved in the biosynthesis of plant products; however, they might also be themselves the producers of many substances of potential use in the modern medicine, agriculture and pharmaceutical industry 5,6 .An area of major interest is to explore endophytic secondary metabolites as novel anticancer and antimicrobial agents.
Since Human African Trypanosomiasis "HAT" (sleeping sickness) is an endemic disease in thirty African countries with population at risk being about 60 million.It caused around 9000 deaths per year 2010, down from 34000 in 1999, which was motivated us to search for significant antitrypanosomal drugs.The infection of HAT is caused by protozoan parasites belong to the genus Trypanosoma which are transmitted to humans by Tsetse fly (Glossina genus) bites 7 .It has two stages: the first one is the haemolymphatic stage which lasts for one to three weeks, followed by the chronic stage in which trypanosomes cross the bloodbrain barrier to invade the central nervous system resulting in chronic meningo-encephalitis and eventually leads to encephalopathy 8 .
Metabolomics is the technique designed to deliver general qualitative and quantitative profiles of metabolites in organisms exposed to various conditions.Plants and microorganisms produce many metabolites with different chemistry and bioactivity under stress conditions.Metabolomics gives access to figure out the complex relationships between the endophytes and their host plants which aids to discover novel bioactive natural components 9 .The metabolome is the complete set of small molecules found in a cell, tissue or organism at a certain point in time.Dereplication is the process of testing sample mixtures which in screening showed activity in order to recognize the novel compounds out of the studied active substances.Dereplication was accomplished by employing differential expression analysis softwares like MZmine which involves dictionary of natural products database (DNP) to aid in compound identification 9 .By using combinations of analytical, statistical and dereplication methods, the bioassayguided isolation route is getting shorter and rapid dereplication of known activities is rapidly delivered 10 .

General instruments
1 H, 13 C and 2D-NMR were recorded at 25ᵒC in DMSO-d6 using Nuclear Magnetic Resonance spectroscopy machine JNM-LA400 model (JEOL, Japan) and the magnet NMR AS400 model EUR0034 (Oxford) Instruments, England at Strathclyde Institute of Pharmacy and Biomedical Science and an AVANCE-III 600 instrument with a 14.1 T Bruker Ultra Shield magnet at Chemistry Department, Faculty of Science, Strathclyde University.ESI-HRMS was measured using FTHRMS-Finnigan LTQ Orbitrap and Exactive mass spectrometer (Thermo Scientific).HPLC analysis was carried out using DionexUltiMate 3000-Thermo Scientific Exactive system instrument, Germany.Crude extracts were initially fractionated using medium pressure liquid chromatography (MPLC) (BÜCHI, Germany), MPLC instrument was the Sepacore Purification System with Versaflash column stand.The Reveleris® Flash Forward system of Grace Davison Discovery Sciences (Illinois, United States) was also used for further isolation, which is characterized of having two detectors, an evaporative light scattering detector (ELSD) and a UV detector (wavelength range: 200-500 nm).The fractions were investigated on normal phase thin layer chromatography plates (TLC silica gel 60 F254), reverse phase TLC plates (TLC silica gel 60 RP-18 F254S) and fractionated using preparative TLC plates (TLC silica gel 60 F254 on 20x20 cm aluminium sheets) (Merck KGaA, Germany).Spots were visualized under UV lamp (short λ nm 254 and long λ nm 380) and after spraying with anisaldehyde and heating till the colour development.
LC-MS spectra were viewed using ThermoXcalibur 2.1 (Thermo Scientific, Germany).To convert the raw data into separate positive and negative ionization files, a program called MS convert from Proteo Wizard was used.The files were then imported to the data mining software MZmine 2.10 for peak picking, deconvolution, deisotoping, alignment and formula prediction 9. Macro file with built in databases was written in Excel, used to combine positive and negative MS files and for further clean-up of media components 11 .The databases used for the identification of compounds were the Dictionary of Natural Products (DNP) 2015, MestReNova (MNova) 2.10 by Mestrelab Research, S.L, (Santiago de Compostela, Spain) was used to process all NMR data and SIMCA 14 (Umetrics AB, Umeå, Sweden) was used for multivariate data analysis.Laminar flow hood (BioMAT2) (Medical Air Technology, UK), the stand incubator (Incu-160S) ,used for agar plates, (SciQuip Ltd., Shropshire), the homogenizer (IKA T18 Basic Ultra-Turrax) and handheld homogenizer (Ultra-Turrax T8) (IKA Labortechnik, Germany) were used for microbiological part.

Plant material
Fresh plant (Lagerstroemia tomentosa) was collected from El-Zohria Botanical Garden in Giza, Egypt and identified by Dr. Therese L. Yousef, senior taxonomist at Orman Garden and by Engineer Mervat A. Hasan, herbarium curator at Orman Botanical Garden.

Isolation of the endophytes
The leaves and stems of the plant were rinsed with sterilized distilled water twice.Surface sterilization was carried out by immersing leaves and stems in 70% isopropanol for 2 min (twice) followed by rinsing again with sterilized distilled water (twice).That was done in order to eliminate surface contaminating microbes.With a sterile scalpel a small segment of leaves and stem tissue (1 cm in length) was cleaned from outer tissue, the inner tissues were carefully dissected under sterile conditions and placed on malt agar plate containing antibiotic to suppress bacterial growth (medium composition: 15 gm agar (Oxoid), 15 gm malt extract (Oxoid) and chloramphenicol (Acros organics, purity > 98%) in distilled water, pH was adjusted to (7.4-7.8) and incubated at 30ºC.After 3-4 weeks, hyphal tips of the fungi were removed and transferred to fresh malt agar medium.Plates were prepared in duplicates to eliminate the possibility of contamination.Pure strains were isolated by repeated inoculation.The purified fungus was later transferred to the liquid medium for scaling up 12 .

Identification of fungal strain
The fungus was isolated as Asperagellus chevellieri according to molecular biological procedure by DNA extraction, amplification and sequencing of the ITS region15.BLAST search of the FASTA sequence as per ormed th the opt on "nr" n lud ng GenBan Ref Seq Nucleotides, EMBL, DDBJ and PDB sequences on the BLAST homepage, NCBI, Bethesda, USA.The accession number was LT670923.
Five different endophytic fungi isolated from different parts of L. tomentosa were identified as Phoma Sp., Chaetomium bostrychodes, Alternaria alternata (from the leaves), Medurella fahi and Aspergillus chevalieri (from the stems).A. chevalieri was selected as it was the most interesting being of unique chemical nature in the dereplication and metabolomic study among all endophytic extracts as well as being biologically active as anti-cancer against ovarian cancer A278O cell line.

Small-scale extraction for screening, metabolomics profiling and dereplication
A plate of each fungal species was transferred into 250 mL flask, then macerated with ethyl acetate and left overnight followed by homogenization and filtration.The mycelia were macerated rice with 200 ml ethyl acetate and filtered.The filtrate was then dried under vacuum.One mg of each extract was subjected to HRMS analysis and 8-10 mg to NMR analysis for metabolomics profiling and dereplication studies.A sample of 1 mg/mL concentration of each fungal extract was prepared in duplicate and was subjected to bioassay against ovarian cancer (A278O), lung cancer (A549), prostatic cancer (PC3) and breast Cancer (ZR75) cell lines.

Fermentation, extraction and isolation
Fresh fungal cultures were transferred into Erlenmeyer flasks (1L each) containing 100 g rice for solid cultures, then incubated at room temperature as stand cultures for 30 days.That was repeated several times to optimize the perfect growth conditions (culture type and incubation period).Moreover, large-scale cultivation was carried out using 20 one-L Erlenmeyer flasks for rice cultures, then 250 ml EtOAc was added to Erlenmeyer flasks and left overnight to stop cell growth.Culture media and mycelia were then homogenized in the Ultraturrax for 10 min for cell destruction, followed by filtration using a Buchner funnel.The mycelium residue was discarded while EtOAc culture filtrates were collected, pooled, dried under vacuum 13 .The dried ethyl acetate extract (10gm) was dissolved in 10% aqueous MeOH and partitioned in a separating funnel with nhexane as a defatting step.The n-hexane (8.2gm) soluble layer was concentrated and subjected to further fractionation a Grace flash chromatography instrument using silica gel cartridge 80g with n-hexane and EtOAc gradient elution to yield compound 1, compound 2 and compound 3.The MeOH soluble portion (880 mg) of the EtOAc extract was then dried and applied to a Grace flash chromatography instrument using silica gel

Cytotoxic activity
Cells were seeded in clear 96 flat-bottomed plates and allowed to adhere overnight.After that time, metabolite extracts and fractions were added at a final concentration of 30 µg/mL, and at a concentration of 10mM/ml for the pure compounds and allowed to incubate for 48 hours.Viability was determined using Alamar Blue® (Thermo Fisher, Paisley, UK), according to the manu a turer's nstru t ons and n u ated or a further 6 h.The resulting fluorescence was measured using a Wallac Victor 2 1420 multi-label counter (Perkin Elmer, Beaconsfield, UK), in fluorescence mode: excitation 560, emission 590.Vehicle treated control cells (media with 0.3% DMSO) were considered 100% viable against which metabolite extract treated cells (at a concentration of 30 µg/mL, at least n = 2) were compared.All results were confirmed microscopically 14 .

Antitrypanosomal activity
Antitrypanosomal activity was tested following the protocol of Huber and Koella 15 .Trypanosoma brucei brucei strain TC 221 were cultivated in Complete Baltz Medium with concentration of 10 4 Trypanosomes per ml.Trypanosomes were tested in 96-well plate chambers against different concentrations of test substances at 0.25-50 µM in 1% DMSO to a final volume of 200 µL.For controls, 1% DMSO as well as parasites without any test compound were used simultanously in each plate to show that DMSO did not perturb the results.The plates were then incubated at 37 °C in an atmosphere of 5% CO 2 for 24 h.After addition of 20 µL of Alamar Blue, the activity was measured after 48 and 72 h by light absorption using an MR 700 Microplate Reader (Dynatech, Chantilly, United States) at a wavelength of 550 nm with a reference wavelength of 650 nm The IC 50 values of the test compound were quantified by linear interpolation of three independent measurements.

Anti-microbial activity
Evaluation of Antimicrobial activity was carried out using a modified Kirby-Bauer disk diffusion assay 16,17 , against pathogenic bacteria (Staphylococcus aureus strain 12600 and Escherichia coli strain 11775) and Fungi (Candida albicans strain 7102).Standard discs of Ampicillin (Antibacterial agent), Amphotericin B (Antifungal agent) served as positive controls for antimicrobial activity and a filter discs impregnated with 10 µl of solvent (DMSO) were used as a negative control.

DISCUSSION
ESI-MS data produced by Excel-macro database file after combining positive and negative modes and removing the media effect, was subjected to R software to apply the heatmap script.The heat map of all extracts showed that Aspergillus chevalieri extract and Alternaria alternate fungal extracts were the richest in metabolites of different mass range as shown in figure 1.The cytotoxicity assay showed that Aspergillus chevalieri was active against A278O cell line and non-toxic for the normal cells PNT2A (Figure 2), implying that Aspergillus chevalieri could have a unique chemical and biological fingerprints.A. chevalieri was then cultivated on small scale solid and liquid cultures to test the optimum growth condition producing the highest amount of interesting metabolites.HRESI-MS data of crude extracts of both rice (RC) and liquid (LC) culture media of A. Chevalieri have been subjected to a metabolomics work flow starting with data mining by MZmine.The heatmap for the processed ESI-MS data of both RC and LC extracts of A. Chevalieri showed more abundance of metabolites in the 30 days RC extract (Figure 3).Moreover, Multivariate data analysis (MVDA) of different culture extracts of A. chevalieri, performed by SIMCA-P V.14 software, discriminated 30-days RC extracts from other fungal extracts as shown in the PCA score plot (Figure 4a) which was indicative of the unique nature of the metabolites produced in RC-30 extract.PCA loading plot (Figure 4b) illustrated the metabolites which could be contributed to the variation of 30 days RC extract.The thirty-day liquid culture extract of A. chevalieri was subjected to fractionation using MPLC, the resulted fractions were imported into SIMCA for MVDA.The bioassay guided MPLC fractionation of the 30-days rice culture extract sorted the active from the inactive fractions according to their activity toward  22,23,24 .This confirmed the efficiency of metabolomics in predicting the bioactive metabolites at first stage of fractionation.However, the rest of significant metabolites in table 3 were not reported to have anticancer activity hence further purification of the active fractions was fundamental to confirm the structure of the previously reported bioactive compound and test the cytotoxicity for the unreported metabolites. aprh.journals.ekb.eg/104

Figure 1 .
Figure 1.Heatmap of ESI-MS data of all endophytic extracts isolated from L. tomentosa (the blue lines represented the produced metabolites).These metabolites were dereplicated by searching DNP 2015 as shown in table 2. Most of these metabolites were reported previously in the literature however metabolites at m/z (retention time in minutes) 581.562 [M+H] + (35.62), 615.528 [M+H] + (35.82), 642.528 [M+H] + (32.95), 767.537 [M-H] -(31.83) and 788.549 [M-H] -(36.02) were not identified in the database.This was motivating for further work on 30 days RC extract, since it was the most active against the selected cancer cell line and showed no toxicity toward normal cells, 30 days RC was chosen for scale up and further isolation work.The thirty-day liquid culture extract of A. chevalieri was subjected to fractionation using MPLC, the resulted fractions were imported into SIMCA for MVDA.The bioassay guided MPLC fractionation of the 30-days rice culture extract sorted the active from the inactive fractions according to their activity toward

Figure 2 .
Figure 2. Cytotoxic activity of all endophytic extract isolated from L. tomentosa.cancer cell line A278O (Figure 7).Fractions have been classified into active and inactive in OPLS-DA analysis.OPLS-DA score plot (Figure 5a) displayed a clustering of fractions 4-5 and 6-24 in the active side.The respective OPLS-DA loading plot (Figure 5b) showed that fraction 4-5 and 6-24 were characterized by these metabolites m/z 153.055 [M+H] + , 291.124 [M-H] -, 301.181 [M-H] -(which were identified in DNP as 2-or 4-hydroxyphenylacetic acid, Curvularin and Auroglaucin; 1',2',3',4'-Tetrahydro respectively.The Splot of active versus inactive fractions showed the most significant metabolites highly correlated to the cytotoxicity of active fractions (Figure 5c).These metabolites were dereplicated as shown in table 3. The bioactivity reported for these metabolites revealed that metabolites at m/z (retention time in minutes) 153.05 [M+H] + (5.99), 291.124 [M-H] -(12.04),301.181 [M-H] - (24.54) equivalent for C 8 H 8 O 3 , C 16 H 20 O 5 and C 19 H 26 O 3 respectively, had cytotoxic and/ or antioxidant activity against different types of cancer cell lines22,23,24 .This confirmed the efficiency of metabolomics in predicting the bioactive metabolites at first stage of fractionation.However, the rest of significant metabolites in table3were not reported to have anticancer activity hence further purification of the active fractions was fundamental to confirm the structure of the previously reported bioactive compound and test the cytotoxicity for the unreported metabolites.

Figure 3 .Figure 4
Figure 3. Heatmap for liquid and rice culture extracts of A. chevalieri (the blue lines represented the produced metabolites).

Figure 5 .
Figure 5. a: OPLS-DA score plot of active versus inactive fractions from 30 days rice culture extract of A. chevalieri, b: OPLS-DA loading plot highlighting the features corresponding to the active fractions, c: S-plot of active versus inactive fractions showing the metabolites highly correlated the anticancer activity of A. chevalieri.