Dereplication of Cytochalasans and Octaketides in Cytotoxic Extracts of Endophytic Fungi from Casearia arborea (Salicaceae)

Endophytes have been shown to be a source of novel drug prototypes. The Casearia genus is known for presenting cytotoxic clerodane diterpenes; however, there are few reports on secondary metabolites produced by its fungal microbiota. Thus, in the present study endophytic fungi obtained from the fresh leaves of C. arborea were grown in potato dextrose broth and rice to perform a secondary metabolite prospection study. The cytotoxic profile of the crude extracts at 10 µg/mL was determined by a colorimetric assay on tumor cell lines. The endophytes producing cytotoxic extracts were identified through phylogenetic analysis and belong to Diaporthe and Colletotrichum species. Metabolites present in these extracts were organized in molecular networking format based on HRMS-MS, and a dereplication process was performed to target compounds for chromatographic purification. Metabolic classes, such as lipids, peptides, alkaloids, and polyketides were annotated, and octaketide and cytochalasin derivatives were investigated. Cytochalasin H was purified from the cytotoxic Diaporthe sp. CarGL8 extract and its cytotoxic activity was determined on human cancer cell lines A549, MCF-7, and HepG2. The data collected in the present study showed that molecular networking is useful to understand the chemical profile of complex matrices to target compounds, minimizing the cost and time spent in purification processes.


Introduction
The Casearia genus (Salicaceae) comprises species with pharmacological properties, which have been used in traditional medicine. C. sylvestris displays anti-inflammatory, antiulcer, anti-ophidian, and antitumor activities [1,2]. Highly oxidized clerodane diterpenes are very known and investigated in Casearia due to their high cytotoxic effects against several tumor cell lines representing promising anticancer natural prototypes [3][4][5].
Secondary metabolites from endophytes, especially those produced by fungi, may be useful as anticancer agents [6][7][8]. The endophytic microorganisms establish interactions with plant hosts and live inside their internal tissues, apparently causing no harm to the host as well as contributing to plant development [9][10][11] Diaporthe spp. is currently associated as a plant pathogen that has been extensively investigated due to its ability to produce several secondary metabolites with biological activities including antibacterial, anticancer, antifungal, antimalarial, antiviral, and herbicidal [12]. The Colletotrichum genus is responsible for causing plant diseases, such as anthracnose, in crops worldwide [13][14][15][16]; however, it may be found as non-pathogen endophytes [17]. Besides the pathogenicity and phytotoxic secondary metabolites, Colletotrichum has been placed among the top 10 fungi genera with economic and scientific importance [18].
Until now, a few endophytic fungi from Casearia species were investigated regarding their metabolites. These studies include alkaloids and polyketides derivatives from Colletotrichum crassipes, as well as diketopiperazines and simple phenolics from Xylaria sp., both fungi isolated from C. sylvestris [19], and a diterpenoid isolated from Neosartorya fischeri, an endophyte from C. grewiifolia [20]. Additionally, Phomopsis sp. from C. arborea produced octaketides, known as cytosporones and dothiorelones [21]. The Casearia mycobiota is still unknown and represents a source to identify fungal metabolites of biological interests. Herein, we report the cytotoxic potential of Casearia arborea endophytes in a colorimetric assay to guide liquid chromatography and mass spectrometry analysis. The online platform Global Natural Products Social Molecular Networking (GNPS) was chosen as a modern tool for spectral organization of metabolites detected in cytotoxic extracts obtained from the incubation of Colletotrichum sp., Diaporthe spp., and Phomopsis sp. to perform the molecular networking (MN). Modern tools, such as MN facilitate annotation processes to aid the chemical profile identification of complex matrixes. Bio-guided assays and MN help to target compounds for chromatographic purification, as described in this work for cytochalasin H from a cytotoxic Diaporthe sp. extract incubation.

Plant Material
Leaves of Casearia arborea Rich. were collected in the Atlantic Forest, Alfenas city, MG, Brazil (coordinates S 21 • 22 53.8" W 045 • 55 46.4"), in June 2016. The botanical identification was performed by João Pedro Costa Elias from the Federal University of Alfenas (UNIFAL-Alfenas city, MG, Brazil), and a voucher specimen (Elias J.P.C. 02) was deposited in the SPF herbarium, São Paulo University.

Isolation of Endophytic Fungi
The methodology used for the isolation of endophytic fungi was based on the procedure described in the literature [22], as well as extraction procedures from liquid and solid cultures [23]. Briefly, healthy leaves were washed in running water and dipped successively into solutions of 70% ethanol (3 min), 1% hypochlorite (1 min), 70% ethanol. After the asepsis, the plant material was transferred to a laminar flow cabinet (Pachane, class Pa40) and dipped in ultra-pure water to remove hypochlorite and alcohol residues. The washed leaves were cut (1-3 mm 2 ) and inoculation was performed in potato dextrose agar (PDA-Kasvi ® ) containing ampicillin (50 mgL −1 ) for fungi growth. The Petri dishes were sealed and the incubation was performed in an incubator at 25 • C. Each morphologically different mycelium that grew out from the plant fragment was subcultured to a plate with PDA medium for the isolation of strains codified as CarGL.

Identification of Endophytic Fungi
The mycelium grown on PDA was ground in liquid nitrogen. The genomic DNA was extracted using the "Wizard ® Genomic DNA Purification Kit" (Promega, Madison, WI, USA). The quality of the reaction was evaluated on agarose gel (1% w/v) with "SYBR ® Safe DNA Gel Stain" (Thermo Fisher Scientific, Waltham, MA, USA). The ITS1-5.8S-ITS2 region of ribosomal DNA was amplified with the universal primers for the Fungi kingdom, ITS-1 (5 TCCGTAGGTGAACCTGCGG-3 ) and ITS-4 (5 TCCTCCGCTTATTGATATGC-3 ), which amplify a region of approximately 600 bp. The reactions were prepared in a final volume of 50 µL containing genomic DNA, 1X enzyme buffer, 0.2 mM dNTPs, 0.2 µM primer ITS-1, 0.2 µM primer ITS-4 and 2.5 U/µL of Easy ® Taq DNA polymerase. The thermal cycler was programmed for an initial denaturation of 5 min at 94 • C, followed by 30 cycles of 94 • C for 30 s; 30 s at 55 • C; 1 min at 72 • C and a final extension of 10 min at 72 • C in a thermal cycler (Veriti 96-Well Thermal Cycler, 0.2 ML, Applied Biosystems, Waltham, MA, USA) [24]. Based on the ITS tree, the isolated fungi from C. arborea clustered together with fourteen species of Diaporthe, in addition to one Colletotrichum sp. The identification was performed using GenBank, Mycobank and BOLD databases (GenBank accession no. MT893333). A Diaporthaceae species (Diaporthella corylina) was used as an external group to construct a phylogenetic tree. The strains CarGL21, 39 and 42 were identified as Diaporthe paranensis. Twelve endophytes were assigned as Diaporthe sp. to strains CarGL2, 5,8,12,18,19,30,31,35,36,37,46, including strain Phomopsis sp. CarGL23 (Supplementary Materials SI), once Phomopsis is the asexual state of Diaporthe comprising hundreds of species belonging to Diaporthaceae [12].

Secondary Fungal Metabolite Prospection
Two culture media were used for endophytic fungi growth: potato dextrose broth (PDB-Kasvi ® , liquid medium) and rice ("Uncle Ben's"-parboiled, solid medium). For PDB incubation five fragments of each isolated fungus mycelium (3 mm 2 ) from PDA incubation were inoculated in Erlenmeyer flasks (500 mL) containing 200 mL of PDB medium previously sterilized. Cultures were maintained on growth at 28 • C on a rotary shaker (120 rpm). The rice medium was prepared in Erlenmeyer flasks (500 mL) containing 50 g of rice and 90 mL of ultra-pure water followed by sterilization (121 • C, 20 min). After 5 days of incubation in PDB 5 mL was transferred to rice for solid incubation at 28 • C under static conditions. After 28 days both fermentations, in rice and PDB, were stopped by adding 200 mL of EtOAc to an Erlenmeyer flask and 20 min in an ultrasonic bath. Extraction was completed under starring on a shaker (120 rpm) for 24 h and then filtered. The extraction procedure was repeated two times. The filtrate was dried in a rotatory evaporator under low pressure at 40-50 • C furnishing the crude EtOAc extracts.

UPLC-ESI-HRMS-MS Data Acquisition
Liquid chromatography coupled to high-resolution tandem mass spectra (LC-HRMS-MS) data were acquired on Shimadzu Nexera X2 ultra-performance liquid chromatography system (Shimadzu, Japan) equipped with an SPD-M20A Proeminence Diode Array detector, using a reverse phase Kinetex EVO C18 column (2.6 µm-100.0 mm × 2.1 mm). All solvents were spectroscopic grade. The LC system was coupled to a QTOF mass spectrometer equipped with an electrospray (ESI) operating in positive ion mode at 18,000 FWHM of mass resolution (MicrOTOF-QII; Bruker Daltonics, Billerica, MA, USA).
Dried EtOAc crude extracts (5 mg) were dissolved in 1 mL of MeOH and centrifuged at 15,000 rpm for 10 min, 20 • C. The supernatant (500 µL) was transferred to a vial and the same volume of MeOH:H 2 O (1:1) was added. Compounds (1 mg mL −1 ) previously obtained from Phomopsis sp. CarGL23 [21] were used to guide the cytosporone and dothiorelone derivatives annotation in the molecular network. Samples were injected (3 µL) into the LC system at 50 • C in a column chamber. The chromatographic separation was performed using the mobile phase in a gradient of A (H 2 O + 0.1% formic acid) and B (ACN + 0.1% formic acid), and a flow of 350 µL min −1 for the following method: 0-2 min 5% B, 2-13 min 5% to 98% B, 13-16 min 98% B, 16-18 min 98% to 5% B, 18-21 min 5% B for the column stabilization for the next injection. HRMS-MS data were obtained from a quadrupole tandem time-of-flight (QTOF) mass analyzer under positive mode ESI at a mass range of m/z 50-1200. The positive ionization on ESI was set as follows: capillarity voltage of 4500 V and end plate offset at 500 V, dry gas (N 2 ) at a flow of 8.0 mL min −1 , a pressure of 4.0 Bar, and temperature of 200 • C. The collision-induced dissociation (CID) energy was set at 25 eV and auto-MS-MS were performed for three precursor ions, with active exclusion after three spectra and release after 1 min, reconsidering the precursor if the current intensity is five times more intense than the previous intensity. Calibration was set at less than 2 ppm using sodium formate.

HRMS-MS Data Organization
All data obtained from Bruker micrOTOF-QII were converted to the ".mzML" extension, to perform the dereplication on GNPS for database matches, using MSConvert [25] and conferred in SeeMS, both are free software from Proteowizard ® . The ".mzML" archives were uploaded to the Global Natural Products Social Molecular Network Web server using WinSCP to create the molecular network [26]. The data were treated within the GNPS Data Analysis platform removing fragments of ±17 Da of precursor m/z. HRMS-MS spectra were filtered choosing only the six top fragments in the ±50 Da window in all ranges of spectra. In the basic options, the mass tolerance for precursor ions and fragment ions was set to 0.02 Da. A network was created using the MS-Cluster with a minimum cluster size containing two spectra [27] according to a cosine score above 0.7 and more than three matched peaks. Further, edges between two nodes were kept in the network if, and only if, each of the nodes appeared in each other's respective top 10 most similar nodes. The maximum size of a molecular family was set to 100, and the lowest scoring edges were removed from molecular families until the molecular family size was below this threshold. The spectra in the network were then searched against GNPS' spectral libraries. The library spectra were filtered in the same manner as the input data. All matches kept between network spectra and library spectra were required to have a score threshold above 0.60 and at least four matched peaks (MSV000086335, doi:10.25345/C5CJ5B). Clusters detected in blank were removed from the networking that includes spectral data from solvents used in the extraction and chromatographic procedures (dried hexane, chloroform, ethyl acetate, methanol). The molecular networking view and edition were performed in Cytoscape v.3.8.2 [28].

Diaporthe sp. CarGL8 Chromatographic Procedures
The extract obtained from Diaporthe sp. CarGL8 incubation in rice furnished a brown and amorphous extract in ethyl acetate which presented cytotoxic effects against three tumor cell lines ( Figure 1). The crude extract was partitioned with hexane and MeOH:H 2 O (9:1). The hydromethanolic phase (400 mg) was then subjected to a chromatographic column (CC) over SiO 2 using a gradient in CHCl 3 :MeOH starting at 95:5, increasing the polarity and ending in 100% MeOH furnishing five group fractions (I-V). Group II was selected by the LC-DAD profile for further purification due to the presence of a major compound. Fraction II was subjected to Sephadex LH-20 CC using methanol as the mobile phase furnishing five fractions (II-1 to II-5). The LC-DAD revealed one chromatographic band in group II-1 (90 mg) at 220 nm with a low absorption chromophore UV max 254 nm. The compound II-1 was then analyzed by NMR and HRMS-MS for structural identification.  supplemented with 10% fetal bovine serum (Vitrocell, Campinas, Brazil). Cells were grown in a humidified atmosphere of 95% air, 5% CO 2 , 37 • C. Cell viability was measured by MTS assay using the CellTiter 96 ® Aqueous Non-Radiative Cell Proliferation assay (Promega) according to the manufacturer's instructions. Cells were seeded into a 96-well plate at 1 × 10 4 cells/well (HepG2, MCF-7, HT144) or 5 × 10 3 cells/well (A549). After attachment (24 h), the cultures were treated with fungal crude extracts at 10 µg mL −1 for 48 h to identify the most cytotoxic extracts. In addition, HepG2, MCF-7, and A549 cells were treated with compound II-1 or cisplatin at different concentrations (0.01-180 µM) for 48 h to determine IC 50 values. The samples were analyzed in a spectrophotometric plate reader at 490 nm. Relative viability was determined by comparing the amount of formazan produced by metabolically active cells in the control and treated groups. Experiments were conducted in triplicate. Data are presented as the mean ± standard deviation (SD) of three independent experiments. IC 50 values were determined from nonlinear regression using GraphPad Prism ® (GraphPad Software, Inc., San Diego, CA, USA) [30].

Results
The authors encourage to visit the GNPS platform and re-analyze the MSV000086335 (Supplementary Materials SIX) that contains the HRMS-MS data obtained from this work. Bellow, based on the MN and fragmentation spectra, cytochalasin and octaketides were investigated for the dereplication process to annotate secondary metabolites in the endophytic metabolome.

Cytotoxic Fungi from Casearia arborea Leaves
Forty-seven endophytic fungi were obtained from fresh and healthy leaves of C. arborea. The fungi inoculation was conducted in two media cultures: solid using rice (R) and liquid using potato and dextrose broth (PDB), for 28 days, thus furnishing ninety-four extracts in ethyl acetate. The cytotoxic potential for crude extracts was evaluated in front of four human tumor cell lines (A549, MCF-7, HepG2, and HT144). Criteria for the strain selection was cell viability less than 70% as result (Figure 1) [21] were also included in the MN to corroborate octaketide derivatives annotations.

Dereplication Based on Molecular Networking Organization from HRMS-MS Data
GNPS annotations were effective to furnish a broad view of the cytotoxic metabolome of endophytes Colletotrichum sp., Diaporthe spp., and Phomopsis sp. The MN obtained from fungi metabolome were organized according to annotations by GNPS including PKS (polyketide synthase) pathway compounds as lipids, cytosporones (Csn), dothiorelones (Dot), ralfuranone L, as well as alkaloids from PKS and NRPS pathway, cytochalasins (Cyt). One fungal dipeptide was detected and annotated as cyclo(Phe-Leu) (Supplementary Materials SIII-SVIII).

Cytochalasin Dereplication
A molecular family with nodes presenting characteristic fragment patterns of m/z 398, 416, and 434 suggested loss of water and/or other substituents as methanol or acetic acid that reminds the alkaloid cytochalasin ESI-HRMS-MS spectra [31,32].
The cytochalasin nodes were detected in Diaporthe sp. CarGL2, 8, and 37 extracts (Supplementary Materials SIII and SIV). Based on cytotoxic activity against A549, MCF-7, and HepG2 and the annotation of the m/z 494 cluster as a cytochalasin, Diaporthe sp. CarGL8 active extract was selected for chromatographic procedures. The compound II-A (Cyt-H) was obtained from SiO 2 CC and identified by spectroscopy data according to the literature [29] and ESI-(+)-HRMS-MS. Moreover, a fragment map for II-A (Cyt-H, m/z 494) was built to annotate cytochalasin derivatives according to the fragmentation mechanisms in the literature [31,32].
The dereplication process allowed the annotation of Cytochalasin J (Cyt-J, m/z 452.279), Cytochalasin J 1 (Cyt-J 1 , m/z 466.295), and Dehydroxy Cytochalasin H (DeOH Cyt-H, m/z 476.279). Furthermore, the m/z 120 ion is also a common fragment according to the loss of biosynthetic precursor phenylalanine moiety. In addition, the node m/z 434 is related to a cytochalasin fragment created in the ionization chamber, which corroborates with finding the presence of cytochalasin derivatives in active extracts from strains CarGL2 and 37 ( Figure 2).  The octaketides Cns-A (m/z 295.156) and Csn-B (m/z 323.180), from past work [21], were added to the GNPS MN workflow to corroborate the identity of the molecular family of cytosporone derivatives based on ESI-(+) fragmentation mechanisms [21].

Cytotoxic Activity of Cytochalasin H
Cell cultures (A549, MCF-7, and HepG2) were treated with compound II-1 at different concentrations for 48 h. The dose-response curves are shown in Figure 4. Cisplatin, a cytotoxic antineoplastic agent, was included in this experimental approach as a positive control. The IC 50 values found for cytochalasin H and cisplatin are shown in Table 1.

Discussion
Polyketide derivatives were the major secondary metabolite class found in Diaporthe spp. endophytes from C. arborea, cited by many researchers as a focus for the search for anticancer natural prototypes [33]. It has been reported that Phomopsis, also identified in C. arborea mycobiota, has cytotoxic activity against several cancer cell lines including breast cancer (MDA-MB-231 and MCF-7), non-small cell lung cancer (A549), hepatoma (HepG2), neuroblastoma (SHSY5Y), and hematologic cancer (HL-60, K562, and Raji) [34][35][36]. In addition, a study carried out by Cui et al. [37] showed that Phomopsis displays inhibitory activity on osteoclastogenesis by suppressing RANKL-induced NF-κB activation.
The octaketides also belong to the PKS pathway presenting a resorcinol lipid skeleton, found in the fungal metabolism of Phoma, Cytospora, and Diaporthe. Natural and synthetic cytosporone B derivatives may be promising prototypes for cancer therapy [46]. These compounds act as an agonist for nuclear orphan receptor TR3, which regulate a diversity of biological processes including proliferation, differentiation, cell survival, and apoptosis [47,48]. TR3 is closely associated with several pathological conditions, such as inflammations, fibrosis, and cancer [49][50][51][52], and it has been considered an important target for drug development [53]. No physiological ligands known for TR3, and Cns-B can regulate its activity. Thus, this compound and others structurally related are of great pharmacological interest [54][55][56][57], and the fungi of C. arborea represent a source to explore cytotoxic PKS metabolites for nuclear orphan receptor TR3 assays.
Cytochalasins, alkaloids from hybrid PKS and NRPS pathways, are well known in the metabolism of Diaporthe spp. and Phomopsis spp. [33]. These compounds were also described in the present work and have been detected in Diaporthe sp. CarGL2, 8, and 37 endophyte strains from C. arborea leaves. Analyses using LC-HRMS-MS associated with the molecular network have indicated the presence Cyt-H. Furthermore, the isolation and evaluation of the activity of the pure substance were carried out to determine if the substance responsible for the activity observed for the extract was in fact Cyt-H. Thus, after isolation, assays were performed with Cyt-H, and activity was detected against A549 and MCF-7 strains, cell lines not previously investigated, with values comparable to those of the standard drug cisplatin (Table 1). Additionally, cytochalasin derivatives are described as acting in actin filaments causing cytokinesis inhibition. At high concentrations, cytochalasin is strongly cytotoxic by inducing the loss of cell denucleation [58,59]. Recently, three cytotoxic Cyt-H derivatives (4 -hydroxy-deacetyl-18-deoxycytochalasin H, deacetyl-18-deoxycytochalasin H, and 18-deoxycytochalasin H) were described from the endophyte Trichoderma harzianum. The cytotoxic assays showed that human cell lines L5178Y and A2780 were most sensitive to 18-deoxycytochalasin H with IC 50 values of 0.19 µM and 0.42 µM, respectively [60]. From Eutypella scoparia, PSU-H267 metabolism (obtained from the leaves of Hevea brasiliensis) was identified Scoparasin C, a cytochalasin derivative that was active against Vero cell lines with IC 50 values of 1.19 µM [61]. Li and colleagues (2018) described that alkaloids and nitrogen-containing natural products are mainly found in endophytic fungi and reported strong cytotoxic activities against several tumor cell lines [7]. Alkaloid, peptide, and nitrogen compounds from fungi have been described with a broad range of biological properties including anticancer activity [6,7,62]. The purification of cytotoxic compound II-A (Cyt-H) targeted and guided by HRMS-MS provided a good integration with bio-guided assays.
In conclusion, the IC 50 values of Cyt-H in tumor cell lines were evaluated and led to good expectations of this substance as an antitumor drug prototype. Thus, the work showed that molecular networking is useful to understand the chemical profile of complex matrices enabling the isolation of target compounds, in our case Cyt-H, as well as minimizing cost and time spent in purification processes. In addition, with the complement of tests using a pure compound, it was possible to determine the real antitumor potential of the Diaporthe fungus extract.  For re-analysis the GNPS platform is also accessible (https://gnps.ucsd.edu/ProteoSAFe/static/ gnps-splash.jsp, accessed on 15 September 2022).

Conflicts of Interest:
The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper.