Rapid Screening and Identification of Antitumor Ingredients from the Mangrove Endophytic Fungus Using an Enzyme-Immobilized Magnetic Nanoparticulate System

As a consequence of recent progression in biomedicine and nanotechnology, nanoparticle-based systems have evolved as a new method with extensive applications in responsive therapy, multimodal imaging, drug delivery and natural product separation. Meanwhile, the magnetic nanoparticulate system has aroused great interest for separation and purification because of its excellent magnetic properties. Phospholipase A2 (PLA2) is a highly expressed regulator to promote the growth of various cancers and is an ideal target to treat cancers. In this study, a novel strategy based on ligand–receptor interactions to discover novel PLA2 inhibitors was established, in which PLA2-functionalized Fe3O4@PLGA-PEG-NH2 magnetic nanoparticles were used as a supporting material combined with high-performance liquid chromatography–mass spectrometry, aiming to accelerate the discovery of novel PLA2 inhibitors from natural sources such as mangrove endophytic fungi. Under the optimized ligand fishing conditions, six target compounds were ultimately fished and identified to be cyclic peptides (1–3) and sterols (4–6), which compounds 1, 2 and 4–6 have well-documented cytotoxicities. Compound 3 exerted better inhibitory effect on A549 cells by experiment. In conclusion, PLA2-functionalized Fe3O4@PLGA-PEG-NH2 magnetic nanoparticles-based ligand fishing provided a feasible, selective and effective platform for the efficient screening and identification of antitumor components from natural products.


Introduction
In recent decades, nanoparticle-based systems have evolved as a new system with extensive applications in multiple biomedical applications, including the diagnosis and therapy of various cancers [1][2][3][4]. In the meanwhile, these systems have also been used for separation due to their unique and inherent advantages such as high efficiency of separation and time-saving compared to the conventional bioassay-guided fractionation methods [5,6].
To date, many compounds extracted from natural products have shown considerable antitumor activities [7], although quick extraction and separation of these compounds still remains challenging. This phenomenon has attracted substantial interest in the discovering of novel methods to isolate the active ingredients from natural products. Ligand fishing, based on the target-ligand binding, was designed to attach the targets such as enzymes, membrane proteins to the carrier material, has been recognized as a convenient and efficient way to achieve multitarget or multichannel screening of natural products, coupled with chromatography or mass spectrometry (HPLC, MS or HPLC-MS) [8,9]. Moreover, new materials used in ligand fishing such as the magnetic nanoparticulate system further enhanced targets' separation due to their great surface area and convenient separation process [10]. Recently, emerging evidence proved that ligand fishing based on magnetic nanoparticles (MNPs) has been applied to screen antitumor compounds from natural products [11,12].
Phospholipase A 2 (PLA 2 ) has been recognized as an important drug target for the initiation and progression of certain types of cancers, including liver, lung, prostate, breast cancers and chronic myelocytic leukemia [13][14][15][16][17]. PLA 2 can specifically cut phospholipids at the sn-2 ester bond to release free fatty acids, mainly arachidonic acid (AA) and lysophospholipids (LPLs) [18]. It mediates the growth and proliferation of cancer cells primarily by releasing AA from glycerophospholipids and the corresponding metabolites from the AA. Besides, PLA 2 may also mediate carcinogenesis by releasing LPLs, which can be metabolized into lysophosphatidic acid (LPA) to promote cell growth [19]. Therefore, PLA 2 or inhibition of PLA 2 was regarded as an important target and an effective way to discover drugs against cancers.
Mangrove endophytic fungi, which grow in tropical and subtropical intertidal estuarine zones, are rich in unique and bioactive compounds with potentials as new medicinal agents [20]. Recently, many novel anticancer compounds with unique structures and skeletons have been identified from mangrove fungi [21]. In addition, PLGA is a polymer with certain degree of hydrophobicity to encapsulate Fe 3 O 4 MNPs via hydrophobic interaction, which is a Food and Drug Administration (FDA)-approved material and considered safe because of its excellent biodegradability and biocompatibility [22]. Considering the adverse limitations of Fe 3 O 4 MNPs [23], it is highly desirable to use PLGA-PEG-NH 2 copolymer as a linker between Fe 3 O 4 MNPs and enzyme. In this study, for the first time, we reported a facile method to screen antitumor compounds from the fungal extracts based on a PLA 2 -functionalized Fe 3 O 4 @PLGA-PEG-NH 2 magnetic nanoparticles (PLA 2 -MNPs) system in combination with LC-MS technology. Amino-terminated PLGA-PEG was firstly synthesized followed by the preparation of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs with emulsion, evaporation and immobilization of PLA 2 on the surface of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs for ligand fishing. This new strategy was validated by screening antitumor compounds from the crude extract of endophytic fungus Pseudopithomyces sp. 1512101. The antitumor activities of screened compounds were determined by MTT assay. As a result, six compounds were efficiently isolated and identified and fusaristatin C (3), one of the mentioned compounds, exhibited better inhibitory effect on A549 cells. The study confirmed the suitability of using PLA 2 -MNPs as a tool of ligand fishing to discover antitumor compounds and accelerate the discovery of new drug candidates from natural products.

Characterization of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs
MNPs have been widely applied in chemical and biological research because of easy surface modifications, excellent stability and convenient solid-liquid separation. At present, most MNPs are based on Fe 2 O 3 , Fe 3 O 4 and other iron oxides. However, the surface of MNPs such as Fe 3 O 4 does not have functional groups to be linked with enzymes or proteins. PLGA-PEG-NH 2 , a di-block copolymer often used as a drug carrier, can be easily modified with various chemical groups due to its low toxicity and high biocompatibility. Therefore, in this study, Fe 3 O 4 MNPs were designed to be encapsulated in the PEG-PLGA-NH 2 carriers to allow magnetic separation and further decoration.
PLGA-PEG-NH 2 was synthesized and covered on the surface of Fe 3 O 4 MNPs to yield Fe 3 O 4 @PLGA-PEG-NH 2 MNPs. The interaction between Fe 3 O 4 MNP surface and PLGA-PEG-NH 2 was mainly based on hydrophobic interaction as both Fe 3 O 4 MNPs and PLGA were substantially hydrophobic. The Fe 3 O 4 MNPs tend to hide inside of the hydrophobic core of the PLGA-PEG-NH 2 polymeric aggregates to minimize the entropy in the aqueous dispersion system and, therefore, formed Fe 3 O 4 @PLGA-PEG-NH 2 MNPs. Using DCC/NHS coupling method, activated PLGA-COOH was reacted with excess NH 2 -PEG-NH 2 to afford PLGA-PEG-NH 2 copolymer ( Figure 1A). The synthesized PLGA-PEG-NH 2 was characterized by FTIR and 1 H NMR spectroscopies. FTIR spectrum ( Figure 1B) confirmed the conjugation of NH 2 -PEG-NH 2 to PLGA-COOH, as evidenced by the vibration peaks at 1625 cm −1 (amide C=O) and 1578 cm −1 (NH bond) which were not shown in the spectrum of PLGA-COOH or NH 2 -PEG-NH 2 . The successful conjugation of PLGA-COOH and NH 2 -PEG-NH 2 was confirmed by the 1 H NMR spectroscopy ( Figure 1C). The peaks at 1.54, 5.18, and 4.79 ppm were assigned to the CH 3 , CH 2 and CH protons of PLGA, respectively, and that at 3.60 ppm corresponded to the CH 2 proton of NH 2 -PEG-NH 2 blocks. Figure S1 shows the FTIR spectra of

Optimization of Preparation Conditions for Fe 3 O 4 @PLGA-PEG-NH 2 MNPs
To optimize the preparation conditions for MNPs, the amount of Fe 3 O 4 MNPs, the volume of dichloromethane and effects of preparation methods were systematically investigated.
Fe 3 O 4 @PLGA-PEG-NH 2 MNPs were prepared by single-emulsion (o/w) and multipleemulsion (w/o/w) methods. As shown in Figure S2A (in the Supplementary Materials), when the magnetic field strength reached 3000 Oe, the saturated magnetization of the MNPs prepared by the single-emulsion method (0.32 emu/g) was 2.6-fold to that prepared by the multiple-emulsion method (0.12 emu/g). Thus, the MNPs prepared by the single-emulsion method showed better magnetic properties.
Moreover, the effects of the amount of Fe 3 O 4 MNPs on the particle size, PDI and saturated magnetization of MNPs were also investigated. As the amount of Fe 3 O 4 MNPs increased, the PDI of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs remained mostly unchanged while the particle size slightly increased ( Figure S2B and Figure S2C). Figure S2D shows the In addition, the content of the free amino groups on the surface of MNPs was measured. The amino groups were determined by the ninhydrin assay with the absorbance at 420 nm measured with UV spectroscopy. A calibration curve was plotted based on the results at various PLGA-PEG-NH 2 concentrations (1.6, 1.8, 2.0, 2.2, 2.4 mg/mL), from which the content of amino group was calculated using the following formula: where A represents the UV absorbance of samples, C represents the concentration of amino groups in the samples. Accordingly, when the amount of dichloromethane was 0.5 mL, the concentration of PLGA-PEG-NH 2 on the MNPs was measured and calculated as 2.21 mg/mL.

Characterizations of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs
Fe 3 O 4 @PLGA-PEG-NH 2 MNPs were characterized by TEM, VSM, particle size analyzer and zeta potential analyzer. The TEM image showed spherical morphology of the prepared MNPs ( Figure 2A). Figure 2A showed that the Fe 3 O 4 @PLGA-PEG-NH 2 MNPs were dispersed in the system with an average diameter of approximately 143 nm. The graph also showed that the Fe 3 O 4 nanoparticles were distributed in the core of the Fe 3 O 4 @PLGA-PEG-NH 2 MNPs with different numbers as the black dots, which may cause the particles to be heterogeneous. In addition, due to the air-drying process during the TEM sample preparation, the polymeric outer layer of the Fe 3 O 4 @PLGA-PEG-NH 2 MNPs might shrink and formed an outer dark layer in the TEM graph. Furthermore, the nanoparticles had a coercivity of 6 Oe ( Figure S3A, in the Supplementary Materials) and a remanence of 0.005 emu/g ( Figure S3B, in the Supplementary Materials). The average size of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs was 141 nm, with a PDI of 0.195 ( Figure 2C), demonstrating excellent dispersion of the particles. The zeta potential of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs was measured to be 17.36 mV and considered to enable the nanoparticles to be relatively stable in the dispersed system.

Activity Study of the Immobilized PLA 2
Immobilization of enzyme onto the surface of MNPs has many advantages compared to the enzyme solutions [24]. Notably, immobilized enzymes are more stable and can be reused repeatedly. As far as we know, there are currently no PLA 2 -modified nanoparticulate systems reported for bioactive compounds screening. Enzymes can be immobilized on the surfaces by covalent bonding or non-covalent interactions. In this study, the covalent bonding method was selected because of the higher stability of the formed enzyme-polymer conjugates. The PLA 2 molecules were conjugated onto the surface of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs with glutaraldehyde linkers to bridge the N-terminus of the PLA 2 molecules and the amino groups on the surface of MNPs by robust covalent bonding.
The content of -NH 2 groups on the surface of MNPs and PLA 2 -MNPs were measured by UV method based on the ninhydrin assay. Compared with MNPs, PLA 2 -MNPs showed a lower UV absorbance at 420 nm because the number of -NH 2 groups decreased after the reaction of PLA 2 with -NH 2 located on MNPs. It confirmed that the successful formation of PLA 2 -MNPs conjugates and the immobilization of the enzyme on Fe 3 O 4 @PLGA-PEG-NH 2 MNPs.
It is necessary to measure the activities of the PLA 2 before and after the immobilization to ensure that the conjugated enzymes are still active. Six standard PLA 2 concentrations (0, 1.25, 2.5, 5, 10, 20 U/L) were tested to obtain a calibration curve and a linear regression equation. The enzyme concentrations of the samples were calculated using the following formula: where Y represents the OD of sample and X represents the concentration of enzyme. Three experimental groups were set to determine the concentrations of immobilized PLA 2 : (1) free PLA 2 ; (2) PBS; (3) PLA 2 -MNPs prepared with different amounts of PLA 2 . According to the commercial provider, the enzymatic activity is 1.2-fold that of its concentration. As shown in Figure 3, enzymatic activity and the binding efficiency of immobilized enzyme increased significantly when the concentrations of enzyme increased from 200 U to 600 U and ceased to increase with higher concentrations of the enzyme. Therefore, PLA 2 immobilization was carried out with 600 U PLA 2 in a subsequent immobilization process. When the amount of PLA 2 was 600 U, the concentrations of free PLA 2 and immobilized enzyme were calculated to be 1.934 U/L and 1.52 U/L, respectively, and the corresponding enzyme activity levels were 2.321 U and 1.824 U, respectively. The activity of PLA 2 -MNPs was nearly 80% of that of free enzyme. The results showed that PLA 2 had been successfully immobilized onto Fe 3 O 4 @PLGA-PEG-NH 2 MNPs, in which the enzymatic activity of PLA 2 was mostly maintained and ready for use thereafter.

Validation of the Ligand Fishing Assay
The mixture (a) of dexamethasone, curcumin and tanshinone IIA, the supernatant (b) after ligand fishing, and the ligand (c) fished by PLA 2 -MNPs were obtained and analyzed by HPLC referring to 3.7. Dexamethasone was chosen as a positive control to assess the specificity of the ligand fishing using a combination of PLA 2 -MNPs and LC-MS. The HPLC chromatograms revealed that tanshinone IIA and curcumin (negative controls) did not bind to PLA 2 enzyme as they were washed off gradually in b and did not appear in c at all. On the contrary, dexamethasone was clearly observed in elute c, which showed that dexamethasone was successfully maintained and fished out with the highest amount and confirmed the specificity of the proposed method ( Figure 4).

Ligand Fishing and LC-MS of Ligands
The HPLC chromatograms of original extract of fungus Pseudopithomyces sp. 1512101 (red line) and elution fractions after ligand fishing by PLA 2 (blue line) were compared ( Figure 5). The HPLC chromatogram of eluent fractions after PLA 2 ligand fishing was significantly simpler and clearer than that of the unfished fungal extracts, implying that considerable substances without affinities to PLA 2 were washed off and PLA 2 -MNPs showed efficient separation of target components. The total ion chromatogram of target ligands is also shown in Figure 6.  After separation and purification, six compounds 1-6 ( Figure 5) were obtained from the extract of fungus Pseudopithomyces sp. 1512101. By comparing the retention time (Rt) and MS with fished ligands, apart from the peak "*" which appeared to be a fatty acid compound with extremely low polarity and failed to be isolated and identified, six compounds were obtained and identified as ligand compounds. Their retention times and molecular ion peaks are listed in Table 1. The structures of cyclo-(4-hydroxyl-Pro-Leu) (1) [25], cyclo-(Pro-Val) (2) [26], fusaristatin C (3) [27], ergosterol peroxide (4) [28], ergosterol (5) [29] and cerevisterol (6) [30] ( Figure 7) were elucidated by comparison with previously reported NMR spectroscopy and MS data. Their 1 H and 13 C NMR spectroscopy data are shown in Table S1 (in the Supplementary Materials). Human lung carcinoma cell line A549, bone marrow neuroblastoma cell line SH-SY5Y and cervical cancer cell line HeLa were selected for cytotoxicity experiment as PLA 2 has been recognized as an important drug target for the initiation and progression of these types of cancers. The results of PLA 2 inhibitory assay demonstrated that compound 3 had an inhibitory effect when the concentration was 60 µM and 100 µM (Table S2, in the Supplementary Materials). According to previous literatures, compounds 1, 2 and 4-6 have been reported with various degrees of cytotoxicity against human cancer cells (Table 2). Therefore, only compound 3 was selected to test its inhibitory activities against human cancer cells in this study. Among the in vitro tests of compound 3 against human lung carcinoma cell line A549, bone marrow neuroblastoma cell line SH-SY5Y and cervical cancer cell line HeLa, the MTT assay revealed that A549 cell line was the most sensitive to the compound 3 and its IC 50 was calculated to be around 10.10 µM (Table 2), lower than the reported results of the other compounds, apart from the result of compound 4 against K562 cell line. The results suggested that compound 3 was a promising inhibitor for PLA 2 and a potential lead compound worthy of further antitumor drug research.

Fungal Strain
The fungus 1512101 was isolated from the leaves of mangrove plant Sonneratia caseolaris, which was collected in October 2015 from the Nansha Mangrove Nature Reserve in Guangzhou, China. The fungal strain was identified to be most similar (99%) to that of Pseudopithomyces sp. 1512101 (compared to MF919624.1). The voucher specimen was stored in our laboratory at 4 • C.

Apparatus and Characterization
UV absorptions were measured using a UV-2550 spectrophotometer (Shimadzu, Kyoto, Japan). All surface functional groups were detected in KBr using a Fourier transform infrared (FTIR) spectrometer (Perkin Elmer, Waltham, MA, USA). The morphology of nanomaterials was observed using a JEM-3100F transmission electron microscope (TEM, JEOL Co., Tokyo, Japan). The particle size of nanomaterials was measured using a Malvern Nano S90 particle size analyzer (Malvern, UK). The zeta potential was recorded on a Delsa TM Nano laser nanoparticle analyzer (Beckman Coulter, Inc., Brea, CA, USA). The magnetic properties were tested with a MPMA XL-7 vibrating sample magnetometer (VSM, Quantum Design, CA, USA). Nanoparticles were freeze-dried using an LGJ-10 freeze dryer (Beijing Songyuan Huaxing Technology Development Co., Ltd., Beijing, China). A YM75 vertical autoclave (Shanghai Sanshen Medical Equipment Co., Ltd., Shanghai, China) was used to sterilize culture medium. A SPX-250C constant temperature and humidity incubator (Shanghai Boxun Industrial Co., Ltd. Medical Equipment Factory, Shanghai, China) and TCYQ shaker (Taicang Experimental Equipment Factory, China) were used for culturing strains. The inhibitory activities of ligands were determined using a Multiskan FC microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). HPLC-MS, carrying out on an Agilent 6120 LC system (Agilent Technologies, Santa Clara, CA, USA), was equipped with a 6520 quadrupole-time-of-flight mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) using an ESI ion source. A Bruker Avance III 400 NMR spectrometer (Bruker, Germany) was employed to obtain the 1 H NMR (400 MHz) and 13 C NMR (100 MHz) spectra of the samples using TMS as an internal standard. A Micromass LCT mass spectrometer (Waters, USA) was used to measure the ESI-TOF mass spectra of the sample, and the accurate mass was determined and calibrated using a lock mass setup. Cells were cultured in an HER Acell 150i incubator (Thermo Fisher Scientific, Waltham, MA, USA).

Preparation of PLGA-PEG-NH 2 di-block Copolymer
PLGA-PEG-NH 2 di-block copolymer was synthesized with the formation of amide linkage between NH 2 -PEG-NH 2 and activated PLGA-COOH as described in previous studies [37][38][39][40][41]. PLGA-COOH (100 mg, 0.01 mmol), 5 molar excesses of NHS (5.6 mg, 0.05 mmol) and DCC (10.3 mg, 0.05 mmol) were dissolved in anhydrous dichloromethane (DCM). The mixture was stirred under N 2 at room temperature for 24 h to activate the carboxyl group of PLGA-COOH. The resulted solution was filtered to remove the formed dicyclohexylurea. The activated PLGA was added dropwise to excess amount of NH 2 -PEG-NH 2 (100 mg, 0.05 mmol) dissolved in anhydrous DCM with gentle stirring. The stoichiometry of PLGA-COOH/NH 2 -PEG-NH 2 was 1:5. The mixture was allowed to be reacted for another 24 h under room temperature. The DCC and NHS were removed from the concentrated solution by precipitation in ice-cold diethyl ether. The amine-terminated PEG-PLGA-NH 2 di-block copolymer was further purified in excess-cold methanol and lyophilized to afford the final product.

Preparation of Fe 3 O 4 @PLGA-PEG-NH 2 MNPs
Fe 3 O 4 @PLGA-PEG-NH 2 MNPs were prepared with a classical solvent emulsion/ evaporation method [42,43]. Briefly, 800 mg of Fe 3 O 4 MNPs and 100 mg of PLGA-PEG-NH 2 were suspended in 0.5 mL of dichloromethane by vortex, giving an oil phase that was then added to 2 mL of 3 wt % aqueous PVA solution and sonicated for 2 min. The emulsion was subsequently added to 15 mL of 1 wt % PVA solution, and the organic phase was removed through reduced pressure evaporation. The dispersion was centrifuged at 6000 rpm for 20 min to remove the unencapsulated Fe 3 O 4 MNPs. The prepared Fe 3 O 4 @PLGA-PEG-NH 2 MNPs product was then freeze-dried and stored at 4 • C.
During the synthesis of

Enzyme Activity Assay
The activity of PLA 2 was tested using an ELISA kit. Enzymatic reactions were performed on a 96-well microplate. Free PLA 2 and immobilized PLA 2 were dissolved in PBS (0.01 M, pH 7.7). PLA 2 was preincubated with HRP-conjugate reagent at 37 • C for 1 h, followed by addition of substrate chromogen solution A (trimethoxybenzaldehyde) and chromogen solution B (tetramethylbenzidine), and incubation at 37 • C for 15 min till the stopping buffer was added into each well. The optical density (OD) of the sample was measured using a microplate reader (450 nm). A standard curve of the samples was plotted using the averaged OD values of six standard concentrations. Finally, the concentration of PLA 2 was determined by comparing OD value of the sample to the standard curve.

Establishment and Validation of Ligand Fishing Assay
Dexamethasone, a typical PLA 2 inhibitor, was chosen as the positive control for the verification of the proposed method. Dexamethasone, curcumin and tanshinone IIA (2 mM each, equimolar) were mixed in PBS as a model sample (a). Three mL of the model sample were added to PLA 2 -MNPs dispersion and incubated for 3 h at 50 • C. After magnetic separation, the supernatant (b) was collected and the MNPs were washed 3 times with PBS. One mL of methanol was used to redisperse the PLA 2 -MNPs for 1 h to dissociate bound components. After separation, the supernatant (c), the mixture a and the supernatant b were collected and analyzed by HPLC after filtration. The column temperature was set at 25 • C. The eluent flow rate was 0.5 mL/min. The mobile phase consisted of solvent A (methanol) and solvent B (0.1%, v/v, formic acid/water). The gradient elution program was optimized as below: 50-85% A at 0-10 min, 85% A at 10-25 min, 85-90% A at 25-30 min, 90% A at 30-40 min. The sample injection volume was 20.0 µL.

Preparation of Fungal Extract
The fresh mycelia of Pseudopithomyces sp. 1512101 were grown on PDA medium at 28 • C for 3-4 days and then inoculated into conical flasks (500 mL) containing 250 mL of PYG medium (10 g/L glucose, 2 g/L peptone, 1 g/L yeast extract, 2 g/L sea salt, pH 6.0-7.0). After incubation for 3-4 days at 28 • C on a rotary shaker at 120 rpm, 5 mL of the culture medium was transferred as the seed into 500 mL flasks containing PYG medium (250 mL). The conical flasks were then incubated for a month at room temperature (25-30 • C). Subsequently, the whole fermented cultures (150 L) were filtered through cheesecloth. The mycelia were separated to obtain a culture broth. Ethyl acetate was added to the culture broth for extraction until the upper layer became colorless, while the mycelia were extracted 3 times with methanol and concentrated. Subsequently, the concentrate was further extracted with ethyl acetate and concentrated to obtain the crude extract (38.4 g).

Application of Ligand Fishing in Fungal Extract
The above fungal extract (3 mL, 1 mg/mL) was incubated at 50 • C with PLA 2 -MNPs for 3 h. After the separation process, PLA 2 -MNPs were taken out and washed 3 times with PBS to remove the components with no or low affinity to PLA 2 followed by incubation in methanol (1 mL) for 1 h to dissociate the specifically bound components. The methanol solution containing potential ligands was collected.

Analysis of Ligands by LC-MS
LC-MS technology was used to analyze the obtained ligands. A Phenomenex C18 column (250 mm × 4.6 mm, 4.5 µm) was used for chromatographic separation. The column temperature was maintained at 25 • C. The eluent flow rate was set at 1 mL/min, and the mobile phase consisted of solvent A (methanol) and solvent B (0.1%, v/v, formic acid/water). The gradient elution program was optimized as below: 50% A at 0-5 min, 50-90% A at 5-45 min, 90-100% A at 45-60 min and 100% A at 60-70 min. The injection volume of the sample was set as 20.0 µL. MS data were acquired in the positive/negative ion mode by ESI and analyzed by PeakView 1.2 ® Software.

In Vitro Cytotoxicity Assay
Compound 3 was evaluated for its inhibitory activities against PLA 2 using a sPLA 2 inhibitor screening assay kit firstly. Then, the cytotoxicities of compound 3 against human lung carcinoma cell line A549, bone marrow neuroblastoma cell line SH-SY5Y and cervical cancer cell line HeLa were determined by the MTT assay. The cells were subcultured when 80-90% confluence was reached, seeded in a 96-well microplate at the density of 5 × 104/mL per well and incubated with various concentrations of compound 3 for 24 h. After, 5-FU was used as the positive control. The cells were treated with 10 µL of MTT in PBS (5 mg/mL) and then incubated for another 4 h. The cells were dissolved in 100 µL of dimethyl sulfoxide. Finally, the OD of the sample at 490 nm was measured and recorded using a microplate reader.

Conclusions
In summary, a new strategy using ligand fishing based on PLA 2 -MNPs with LC-MS was established to separate and analyze bioactive components from natural products. Using this strategy, six ligands of PLA 2 were rapidly extracted and identified from mangrove endophytic fungi. Furthermore, the cytotoxicity of these compounds was also evaluated. This robust and relatively convenient ligand fishing approach using PLA 2 -MNPs can be applied to screen more antitumor ingredients from natural extracts. In the future, we can construct a new nanoparticle system to load targeted enzyme compounds for the treatment of cancer in vivo.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.